* Encoding: ISO-8859-1.
title'indirect effect age (x), experience (m), positive (y)'.

DATASET ACTIVATE DataSet1.
/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = positive age xp 
  /y=positive/x=age/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.

title'indirect effect gender (x), experience (m), positive (y)'.



/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = positive gender xp 
  /y=positive/x=gender/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.

title'indirect effect age (x), experience (m), negative (y)'.




/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = negative age xp 
  /y=negative/x=age/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.

title'indirect effect gender (x), experience (m), negative (y)'.




/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = negative gender xp 
  /y=negative/x=gender/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.

* Encoding: ISO-8859-1.
title'indirect effects of age (x) on attitude (y) by experience (z)'



DATASET ACTIVATE DataSet1.
/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = attitude age xp 
  /y=attitude/x=age/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.

title'indirect effects of gender (x) on attitude (y) by experience (z)'





/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = attitude gender xp 
  /y=attitude/x=gender/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.

title'indirect effects of age (x) on shootings (y) by experience (z)'.







/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = shoot age xp 
  /y=shoot/x=age/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.

title'indirect effects of gender (x) on shooting (y) by experience (z)'.


/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = shoot gender xp 
  /y=shoot/x=gender/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.

title'indirect effects of age (x) on legislation (y) by experience (z)'.




/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = newlaws age xp 
  /y=newlaws/x=age/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
RESTORE.

title'indirect effects of gender (x) on legislation (y) by experience (z)'.



/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = newlaws gender xp 
  /y=newlaws/x=gender/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.


* Encoding: ISO-8859-1.


title'indirect effect of age (x) on accurate positive (y) by experience (z)'.


/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = eap age xp 
  /y=eap/x=age/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.

title'indirect effect of caregiver (x) on accurate positive (y) by experience (z)'.



/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = eap parent xp 
  /y=eap/x=parent/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.

title'indirect effect of co-play (x) on accurate positive (y) by experience (z)'.





/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = eap coplay xp 
  /y=eap/x=coplay/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.

title'indirect effect of age (x) on accurate negative (y) by experience (z)'.



/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = ean age xp 
  /y=ean/x=age/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.

title'indirect effect of gender (x) on accurate negative (y) by experience (z)'.




/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = ean gender xp 
  /y=ean/x=gender/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.

title'indirect effect of caregiver (x) on accurate negative (y) by experience (z)'.



/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = ean parent xp 
  /y=ean/x=parent/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.

title'indirect effect of caregiver coplay (x) on accurate negative (y) by experience (z)'.



/* PROCESS for SPSS 2.15 */.
/* Written by Andrew F. Hayes */.
/* www.afhayes.com */.
/* Copyright 2012-2016 */.
/* Online distribution other than through */.
/* www.afhayes.com or processmacro.org is not authorized */.
/* Please read the documentation */.
/* available in Appendix A of */.
/* Hayes (2013) prior to use */.
/* www.guilford.com/p/hayes3 */.
/* Documentation available in Appendix A of http://www.guilford.com/p/hayes3 */.
preserve.
set printback=off.
DEFINE hcest (hcdat=!charend('/')/hcres=!charend('/')/
  hctog=!charend('/')/hcmse=!charend('/')).
compute hcinvXtX=inv(t(!hcdat)*!hcdat).
compute k3 = ncol(!hcdat).
do if (!hctog = 1).
  compute h = !hcdat(:,1).
  loop i3=1 to nrow(!hcdat).
    compute h(i3,1)= !hcdat(i3,:)*hcinvXtX*t(!hcdat(i3,:)). 
  end loop.
  loop i3=1 to k3.
    compute !hcdat(:,i3) = (!hcres(:,ncol(!hcres))&/(1-h))&*!hcdat(:,i3).
  end loop.
end if.
do if (!hctog <> 1).
  loop i3=1 to k3.
    compute !hcdat(:,i3) = sqrt(!hcmse)&*!hcdat(:,i3).
  end loop.
end if.
compute hcdat=!hcdat.
compute covmat=hcinvXtx*t(!hcdat)*!hcdat*hcinvXtX.
!ENDDEFINE.
define bcboot (databcbt = !charend ('/')/estmte = !charend ('/') !default(9999)).
compute temp = !databcbt.
compute temp(GRADE(!databcbt)) = !databcbt.
compute badlo = 0.
compute badhi = 0.
do if (!estmte <> 9999).
  compute pv=csum(temp < !estmte)/boot.
  compute ppv = pv. 
  do if (pv > .5).
    compute ppv = 1-pv.
  end if.
  compute y5=sqrt(-2*ln(ppv)).
  compute xp=y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0).
  do if (pv <= .5).
    compute xp = -xp.
  end if.
  compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
  compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
  do if (cilow < 1).
    compute cilow = 1.
     compute booterr=1.
     compute badlo = 1.
  end if.
  do if (cihigh > boot).
    compute cihigh = boot.
    compute booterr=1.
     compute badhi = 1.
  end if.
  compute llcit=temp(cilow,1).
  compute ulcit=temp(cihigh,1).
end if.
do if (!estmte = 9999).
   compute llcit=temp(cilow,1).
   compute ulcit=temp(cihigh,1).
end if.
!enddefine.
define bcci (var = !charend ('/')/point = !charend ('/') !default(9999)/conf = !charend('/') 
    !default(95)).
MATRIX.
get perdata/file = */variables = !var/missing=omit.
compute berror=0.
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute conf=!conf.
do if (trunc(conf) ge 100 or (trunc(conf) le 50)).
  compute conf = 95.
end if.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute boot=nrow(perdata).
compute temp = perdata.
compute temp(GRADE(perdata)) = perdata.
compute badlo = 0.
compute badhi = 0.
do if (!point <> 9999).
  compute pv=csum(temp < !point)/boot.
  do if ((pv = 0) or (pv = 1)).
     compute berror=2.
  end if.
  do if (berror=0).
    compute ppv = pv. 
    do if (pv > .5).
      compute ppv = 1-pv.
    end if.
    compute y5=sqrt(-2*ln(ppv)).
    compute xp=((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0).
    compute xp=(xp/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0))+y5.
    do if (pv <= .5).
      compute xp = -xp.
    end if.
    compute cilow=rnd(boot*(cdfnorm(2*xp+xp2))).
    compute cihigh=trunc(boot*(cdfnorm(2*xp+(-xp2))))+1.
    do if (cilow < 1 or cihigh > boot).
      compute berror=1.
    end if.
    do if (berror=0).
    compute llcit=temp(cilow,1).
    compute ulcit=temp(cihigh,1).
    compute outp={conf; llcit; ulcit}.
    print outp/title = "Bias corrected confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4.
    end if.
  end if.
end if.
do if (!point  = 9999).
  compute cilowp=rnd(boot*(cdfnorm(xp2))).
  compute cihighp=trunc(boot*(cdfnorm(-xp2)))+1.
  do if (cilowp < 1 or cihighp > boot).
     compute berror=1.
  end if.
  do if (berror = 0).
    compute llcitp=temp(cilowp,1).
    compute ulcitp=temp(cihighp,1).
    compute outp ={conf; llcitp; ulcitp}.
    print outp/title = "Percentile confidence interval:"/rlabels "Conf%" "LLCI" "ULCI"/format = 
    F10.4..
  end if.
end if.
do if (berror=1).
  print/title = "Error: Decrease your confidence or increase the number of bootstrap estimates.".
end if.
do if (berror=2).
  print/title = "Error: Impossible point estimate provided.".
end if.
END MATRIX.
!enddefine.
DEFINE LOGIT2 (pt1lp=!charend ('/')/xlp=!charend('/')/ylp=!charend('/')).
compute pt1lp=!pt1lp.
compute xlp=!xlp.
compute ylp=!ylp.
loop jjj = 1 to iterate.
  compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
  compute b = bt1+inv(t(xlp)*vt1*xlp)*t(xlp)*(ylp-pt1lp).
  compute pt1lp = 1/(1+exp(-(xlp*b))).
  compute itprob = csum((pt1lp < .00000000000001) or (pt1lp > .99999999999999)).
  do if (itprob > 0).
    loop kkk = 1 to nrow(pt1lp).
      do if (pt1lp(kkk,1) = 1).
        compute pt1lp(kkk,1) = .99999999999999.
      end if.
      do if (pt1lp(kkk,1) = 0).
        compute pt1lp(kkk,1) = .00000000000001.
      end if.
    end loop.
    compute itprob = 0.
  end if.
  do if (itprob = 0).
    compute LL = ylp&*ln(pt1lp)+(1-ylp)&*ln(1-pt1lp).
    compute LL2 = -2*csum(ll).
  end if.
  do if (abs(LL1-LL2) < converge).
    compute vt1 = mdiag(pt1lp&*(1-pt1lp)).
    compute varb = inv(t(xlp)*vt1*xlp).
    compute seb = sqrt(diag(varb)).
    break.
  end if.
  compute bt1 = b.
  compute LL1 = LL2.
end loop.
!ENDDEFINE.
DEFINE PRCD123 (vars = !charend('/')/model = !charend('/') !default(77)/y = !charend('/')/m = 
    !charend('/')/x = !charend ('/')
   /w = !charend('/') !default(xxx)/z = !charend('/') !default(xxx)/v = !charend('/') 
    !default(xxx)/q = !charend ('/') !default(xxx)/conf = !charend('/') !default(95)
   /hc3 = !charend('/') !default(0)/cluster = !charend('/') !default(xxx)/wmodval = !charend('/') 
    !default (999)/zmodval = !charend('/') !default(999)/
   vmodval = !charend('/') !default (999)/qmodval = !charend('/') !default(999)/mmodval = 
    !charend('/') !default (999)/xmodval = !charend('/') !default (999)
  /boot = !charend('/') !default(1000)/center = !charend('/') !default(0)/quantile = !charend('/') 
    !default(0)/effsize = !charend('/') !default(0)/normal = !charend('/') !default(0)
   /varorder = !charend('/') !default(2)/total = !charend('/') !default(0)/plot = !charend('/') 
    !default(0)/detail = !charend('/') !default(1)
   /iterate = !charend('/') !default(10000)/converge = !charend('/') !default(0.00000001)/percent = 
    !charend('/') !default(0)/jn = !charend('/') !default(0)/
     coeffci = !charend('/') !default(1)/covmy = !charend('/') !default(0)/contrast = !charend('/') 
    !default(0)/seed = !charend('/') !default(random)
      /save = !charend('/') !default(0)/mc= !charend('/') !default(0)/decimals=!charend('/') 
    !default(F10.4)/ws=!charend('/') !default(0)/olsdichm=!charend('/') !default(0)
      /olsdichy=!charend('/') !default(0)/covcoeff=!charend('/') !default(0)/mcx=!charend('/') 
    !default(0)/
      mcm=!charend('/') !default(0)/longname=!charend('/') !default(0)).
set printback = off.
set seed = !seed.
string w999999t (A8).
string z999999t (A8).
string q999999t (A8).
string v999999t (A8).
compute w999999t = (!quote(!w)).
compute z999999t = (!quote(!z)).
compute q999999t = (!quote(!q)).
compute v999999t = (!quote(!v)).
set mxloop = 100000000.
matrix.
get dat/file = */variables = !vars/names = vnames/missing = 9999.
compute ninit = nrow(dat). 
get dat/file = */variables = !vars/names = vnames/missing = omit. 
get tmp/file = */variables = !y/names = yname/missing = omit.
get tmp2/file = */variables = !x/names = xname/missing = omit.
get tmp/file = */variables = !m/names = mnames/missing = omit.
get tmp/file = */variables = w999999t z999999t v999999t q999999t.
compute wname=tmp(1,1).
do if (wname = ' ').
  compute wname = "xxx".
end if.
compute zname=tmp(1,2).
do if (zname = ' ').
  compute zname = "xxx".
end if.
compute vname=tmp(1,3).
do if (vname = ' ').
  compute vname = "xxx".
end if.
compute qname=tmp(1,4).
do if (qname= ' ').
  compute qname = "xxx".
end if.
compute n = nrow(dat).
compute p0=-.322232431088.
compute p1 = -1.
compute p2 = -.342242088547.
compute p3 = -.0204231210245.
compute p4 = -.0000453642210148.
compute q0 = .0993484626060.
compute q1 = .588581570495.
compute q2 = .531103462366.
compute q3 = .103537752850.
compute q4 = .0038560700634.
compute badend = 0.
compute priorlo = -9999999.
compute priorhi = 9999999.
compute criterr = 0.
compute cluster = 0.
compute clsdmy = 0.
compute jndich = 0.
compute wvdich=0.
compute mod74dic=0.
compute warnrep=0.
compute contrast = trunc(!contrast).
do if (contrast > 2 or contrast < 0).
  compute contrast = 0.
end if.
compute booterr = 0.
compute effsize = (!effsize = 1).
compute note = make(10,1,0).
compute notes = 1.
compute iterr = 0.
compute clsmtch = 0.
compute quantile = (!quantile = 1).
compute mmodval=!mmodval.
compute jn = (!jn = 1).
compute covcoeff=(!covcoeff=1).
compute center = (!center = 1).
compute detail = (!detail = 1).
compute coeffci = (!coeffci = 1).
compute conf = !conf.
compute longname = (!longname=1).
compute bconoff=(!percent <> 1).
compute covmy = trunc(!covmy).
do if (covmy < 0 or covmy > 2).
   compute covmy = 0.
end if.
do if (trunc(conf) >= 100 or (trunc(conf) <= 50)).
  compute conf = 95.
  compute note(notes,1) = 1.
  compute notes = notes + 1.
end if.
do if (n < ninit).
  compute nmiss = ninit-n.
  compute note(notes,1) = 11.
  compute notes = notes + 1.
end if.
compute errs = 0.
compute quantd = {0,0,0,0,0,0}.
compute quantc = {0,0,0,0,0,0}.
compute mcheck = 0.
compute ttt = 0.
compute plot = trunc(!plot).
compute runerrs = make(50,1,0).
compute model = trunc(!model).
compute normal = !normal.
compute ws= (!ws=1).
compute olsdichm= (!olsdichm=1).
compute olsdichy= (!olsdichy=1).
compute mcx=trunc(!mcx).
compute mcm=trunc(!mcm).
do if (ws = 1).
  do if (effsize = 1).
    compute note(notes,1) = 19.
    compute notes = notes + 1.
    compute effsize=0.
  end if.
  do if (normal=1).
    compute note(notes,1) = 16.
    compute notes = notes + 1.
    compute normal=0.
  end if.
end if.
do if (model <> 4).
  compute ws = 0.
end if.
print/space=0.
do if (jn = 1 and model <> 1 and model <> 3).
  compute note(notes,1) = 7.
  compute notes = notes + 1.
end if.
do if ((model > 76) or (model < 1)).
  compute model = 77.
  compute criterr = 1.
  compute errs = errs+1.
  compute runerrs(errs,1) = 19.
end if.
compute itprob=0.
do if (mcx < 0).
  compute mcx=0.
end if.
do if (mcx > 4).
  compute mcx=0.
end if.
do if (mcm < 0).
  compute mcm=0.
end if.
do if (mcm > 4).
  compute mcm=0.
end if.
do if (mcx > 0 and mcm > 0).
  compute errs=errs+1.
  compute runerrs(errs,1)=34.
  compute criterr=1.
end if.
do if ((model <> 1 and model <> 4) and (mcx > 0 or mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=35.
  compute criterr=1.
end if.
do if ((model = 4) and (mcm > 0)).
  compute errs=errs+1.
  compute runerrs(errs,1)=41.
  compute criterr=1.
end if.
do if ((mcx > 0) and jn <> 0).
  compute jn=0.
  compute note(notes,1) = 23.
  compute notes = notes + 1.
end if.
do if (mmodval <> 999 and mcm > 0).
  compute note(notes,1) = 24.
  compute notes = notes + 1.
end if.
compute toteff = 0.
compute toteff = ((!total = 1)*(!model = 4 or !model = 6)).
compute varorder = !varorder.
compute hc3 = (!hc3 <> 0).
compute clname = !quote(!cluster).
compute centvar = {"xxx"}.
compute nmeds = ncol(mnames).
do if (longname=0).
  !let !toomany=0.
  !do !i !in (!vars).
    !do !j = 1 !to !length(!i).
      !if ((!j > 8) !and (!toomany = 0)) !then.
        compute criterr = 1.
        compute errs = errs+1.
        compute runerrs(errs,1) = 33.
        !let !toomany = 1.
      !ifend.
    !doend.
  !doend.
end if.
compute modelm =
{0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,2;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,3;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4;
 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,5;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6;
 1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7;
 1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,8;
 1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9;
 1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,10;
 1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11;
 1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,12;
 1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,13;
 0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14;
 0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,15;
 0,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,16;
 0,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,17;
 0,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,18;
 0,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,19;
 0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,20;
 1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,21;
 1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,22;
 1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,23;
 1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,24;
 1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,25;
 1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,26;
 1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,0,0,0,27;
 1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,28;
 1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,29;
 1,1,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,30;
 1,1,0,1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,31;
 1,1,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,32;
 1,1,1,1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,33;
 1,1,1,1,0,0,1,0,0,1,0,0,0,0,0,0,0,0,34;
 1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,35;
 1,0,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,36;
 1,0,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,37;
 1,0,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,38;
 1,0,0,1,1,1,0,0,0,1,0,0,0,0,0,0,0,0,39;
 1,0,0,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,40;
 1,0,0,1,1,0,1,0,0,1,1,0,0,0,0,0,0,0,41;
 1,0,0,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,42;
 1,0,0,1,1,1,1,0,0,1,1,1,0,0,0,0,0,0,43;
 1,0,0,1,1,1,1,0,0,1,0,0,0,0,0,0,0,0,44;
 1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,45;
 1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,46;
 1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,47;
 1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,48;
 1,1,0,1,1,0,1,1,0,0,0,0,0,0,0,0,0,0,49;
 1,1,0,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,50;
 1,1,1,1,1,0,1,1,1,0,0,0,0,0,0,0,0,0,51;
 1,1,1,1,1,0,0,0,0,1,1,0,0,0,0,0,0,0,52;
 1,1,0,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,53;
 1,1,0,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,54;
 1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,55;
 1,1,1,1,1,1,0,0,0,1,1,1,0,0,0,0,0,0,56;
 1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,57;
 1,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,58;
 1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,59;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,60;
 1,1,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,61;
 1,1,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0,0,62;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,0,0,0,63;
 1,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,64;
 1,0,0,1,0,0,1,0,0,0,0,0,1,0,0,0,0,0,65;
 1,0,0,1,0,0,0,0,0,1,0,0,1,0,0,0,0,0,66;
 1,0,0,1,0,0,1,0,0,1,0,0,1,0,0,0,0,0,67;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,68;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,0,0,0,69;
 1,0,0,1,0,0,0,0,0,0,0,0,1,1,0,0,0,0,70;
 1,0,0,1,0,0,1,0,0,1,0,0,1,1,1,0,0,0,71;
 1,1,1,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,72;
 1,1,1,0,0,0,1,1,1,0,0,0,1,0,0,1,1,0,73;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,74;
 1,1,0,0,0,0,0,0,0,0,0,0,1,0,0,1,0,0,75;
 1,1,0,0,0,0,1,1,0,0,0,0,1,0,0,1,0,0,76;
 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,77}.
compute wm = modelm(model, 1).
compute zm = modelm(model, 2).
compute wzm = modelm(model, 3).
compute vy = modelm(model, 4).
compute qy = modelm(model, 5).
compute vqy = modelm(model, 6).
compute wy = modelm(model, 7).
compute zy = modelm(model, 8).
compute wzy = modelm(model, 9).
compute vxy = modelm(model, 10).
compute qxy = modelm(model, 11).
compute vqxy = modelm(model, 12).
compute wmy = modelm(model, 13).
compute wvmy = modelm(model, 14).
compute wvxy = modelm(model, 15).
compute zmy = modelm(model,16).
compute wzmy = modelm(model,17).
compute xmy = modelm(model,18).
do if (ncol(xname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 20.
  compute criterr = 1.
end if.
do if (ncol(mnames) > 10).
  compute errs = errs+1.
  compute runerrs(errs,1) = 39.
  compute criterr = 1.
end if.
do if (ncol(yname) <> 1).
  compute errs = errs+1.
  compute runerrs(errs,1) = 21.
  compute criterr = 1.
end if.
do if ((!mmodval=0) and (!wmodval=0) and (model=3 or model=2) and (contrast=1)).
  compute errs = errs+1.
  compute runerrs(errs,1) = 31.
  compute criterr = 1.
end if.
compute xlist = (wm or zm or wzm or wy or zy or wzy or vxy or qxy or vqxy or wvxy or xmy).
compute mlist = (vy or qy or vqy or zmy or wmy or wzmy or xmy or (model < 4)).
compute bad=0.
do if (criterr = 0).
  compute werr = 0.
  compute verr = 0.
  compute qerr = 0.
  compute zerr = 0.
  compute yerr = 1.
  compute xerr = 1.
  compute wlist = (wm or wzm or wy or wzy or wm or wvmy or wvmy or wvxy or wzmy).
  do if (wlist = 1 and wname = "xxx").
    compute werr = 1.
    compute wlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 4.
  end if.
  do if (wlist = 1 and ((wname = qname) or (wname = vname) or (wname = zname) or (wname = xname) or 
    (wname = yname))).
    compute werr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 12).
  end if.
  compute zlist = (zm or wzm or zy or wzy or zmy or wzmy).
  do if (zlist = 1 and zname = "xxx").
    compute zerr = 1.
    compute zlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (zlist = 1 and ((zname = qname) or (zname = vname) or (zname = wname) or (zname = xname) or 
    (zname = yname))).
    compute zerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 13.
  end if.
  compute qlist = (qy or vqy or qxy or vqxy).
  do if (qlist = 1 and qname = "xxx").
    compute qerr = 1.
    compute qlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (qlist = 1 and ((qname = zname) or (qname = vname) or (qname = wname) or (qname = xname) or 
    (qname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 14.
  end if.
  compute vlist = (vy or vqy or vxy or vqxy or wvmy or wvxy).
  do if (vlist = 1 and vname = "xxx").
    compute  verr = 1.
    compute vlist = 0.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (vlist = 1 and ((vname = zname) or (vname = qname) or (vname = wname) or (vname = xname) or 
    (vname = yname))).
    compute qerr = 4.
    compute errs = errs+1.
    compute runerrs(errs,1) = 15.
  end if.
  do if (wlist = 0 and wname <> "xxx").
    compute werr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 8.
  end if.
  do if (zlist = 0 and zname <> "xxx").
    compute zerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 9.
  end if.
  do if (qlist = 0 and qname <> "xxx").
    compute qerr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 10.
  end if.
  do if (vlist = 0 and vname <> "xxx").
    compute verr = 2.
    compute errs = errs+1.
    compute runerrs(errs,1) = 11.
  end if.
  do if (hc3 = 1).
    compute note(notes,1) = 3.
    compute notes = notes+1.
  end if.
  compute alpha2 = (1-(conf/100))/2.
  compute y5=sqrt(-2*ln(alpha2)).
  compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
  compute cons = make(n,1,1).
  compute temp = (n*sscp(dat))-(t(csum(dat))*(csum(dat))).
  compute temp = temp/(n*(n-1)).
  compute temp = csum(((diag(temp))=0)).
  compute temp2=1.
  do if (ws = 1).
    compute temp2 = (n*sscp(tmp2))-(t(csum(tmp2))*(csum(tmp2))).
    compute temp2 = temp2/(n*(n-1)).
  end if.
  do if ((temp > 0 and ws = 0) or (temp > 0 and ws = 1 and temp2 > 0)).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 27.
  end if.
  compute nmeds = ncol(mnames).
  compute mcmats=ident(nmeds*2).
  compute mccoeff=make(nmeds*2,1,0).
  compute sobel = make(nmeds,4,-999).
  do if (model = 6 and nmeds > 4).
    compute errs = errs+1.
    compute runerrs(errs,1)=2.
  end if.
  do if (model < 4 and nmeds > 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 3.
  end if.
  compute nmods = (model = 74).
  compute bad = 0.
  compute intcnt = 1.
  compute modvals = 0.
  compute modvalsd = 0.
  compute yintemp = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", 
    "int_9", "int_10", "int_11", "int_12", "int_13", "int_14", "int_15"}.
  compute yintemp = {yintemp, "int_16", "int_17", "int_18", "int_19", "int_20", "int_21", "int_22", 
    "int_23", "int_24", "int_25", "int_26", "int_27", "int_28"}.
  compute yintemp = {yintemp, "int_29", "int_30", "int_31", "int_32", "int_33", "int_34", "int_35", 
    "int_36", "int_37", "int_38", "int_39", "int_40"}.
  compute yintemp = {yintemp, "int_41", "int_42", "int_43", "int_44", "int_45", "int_46", "int_47", 
    "int_48", "int_49", "int_50", "int_51", "int_52"}.
  compute yintemp = {yintemp, "int_53", "int_54", "int_55", "int_56", "int_57", "int_58", "int_59", 
    "int_60", "int_61", "int_62", "int_63", "int_64"}.
  compute cntname={"(C1)";"(C2)";"(C3)";"(C4)";"(C5)";"(C6)";"(C7)";"(C8)";"(C9)";"(C10)";"(C11)"+
    "";"(C12)";"(C13)";"(C14)";"(C15)";"(C16)";"(C17)"}.
  compute cntname={cntname;"(C18)";"(C19)";"(C20)";"(C21)";"(C22)";"(C23)";"(C24)";"(C25)";"(C26)"+
    "";"(C27)";"(C28)";"(C29)";"(C30)";"(C31)"}.
  compute cntname={cntname;"(C32)";"(C33)";"(C34)";"(C35)";"(C36)";"(C37)";"(C38)";"(C39)";"(C40)"+
    "";"(C41)";"(C42)";"(C43)";"(C44)";"(C45)"}.
  compute cntname={cntname;"(C46)";"(C47)";"(C48)";"(C49)";"(C50)";"(C51)";"(C52)";"(C53)";"(C54)"+
    "";"(C55)";"(C56)";"(C57)";"(C58)";"(C59)"}.
  compute cntname={cntname;"(C60)";"(C61)";"(C62)";"(C63)";"(C64)";"(C65)";"(C66)";"(C67)";"(C68)"+
    "";"(C69)";"(C70)";"(C71)";"(C72)";"(C73)"}.
  compute cntname={cntname;"(C74)";"(C75)";"(C76)";"(C77)";"(C78)";"(C79)";"(C80)";"(C81)";"(C82)"+
    "";"(C83)";"(C84)";"(C85)";"(C86)";"(C87)"}.
  compute cntname={cntname;"(C88)";"(C89)";"(C90)";"(C91)";"(C92)";"(C93)";"(C94)";"(C95)";"(C96)"+
    "";"(C97)";"(C98)";"(C99)";"(C100)";"(C101)"}.
  compute cntname={cntname;"(C102)";"(C103)";"(C104)";"(C105)"}.
  compute apathnam={"a path"; "a1 path"; "a2 path"; "a3 path"; "a4 path"; "a5 path"; "a6 path"; 
    "a7 path"; "a8 path"; "a9 path"; "a10 path"}.
  compute bpathnam={"b path"; "b1 path"; "b2 path"; "b3 path"; "b4path"; "b5path"; "b6 path"; 
    "b7 path"; "b8 path"; "b9 path"; "b10 path"}.
  compute modvnm = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute modvnm2 = {"xxx", "xxx", "xxx", "xxx", "xxx"}.
  compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 ="; 
    "   M8 ="; "   M9 ="; "  M10 ="}.
  compute m = make(n,nmeds,1).
  compute mmat = make(16,nmeds,0).
  compute ymat = make(8,nmeds,0).
  compute deco = make(10,1,0).
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  compute iterate = abs(trunc(!iterate)).
  compute converge = abs(!converge).
  compute boot = abs(trunc(!boot)).
  do if (ws=1 and !mc > 0).
    compute note(notes,1)=17.
    compute notes=notes+1.
  end if.
  compute mc=(abs(trunc(!mc))*(1-ws)).
  do if (mc > 0 and model > 5).
    do if (boot = 0).
      compute boot=mc.
    end if.
    compute mc=0.
    compute note(notes,1)=12.
    compute notes = notes+1.
  end if.
  compute adjust = 0.
  do if (mc > 0 and (model > 3 and model < 6)).
    compute boot=0.
    compute bconoff=0.
  end if.
  do if (boot > 0 and mc > 0).
    compute mc = 0.
  end if.
  compute savboot = 0.
  compute saveboot = (!save = 1).
  do if (ws=1 and saveboot=1).
    compute saveboot=0.
    compute note(notes,1) = 20.
    compute notes = notes + 1.
    compute normal=0.
  end if.
  do if (saveboot =1 and boot > 0 and model > 3).
    compute savboot = 1.
  end if.
  do if (boot <> 0 or mc <> 0).
    compute bootsz=boot.
    do if (mc > 0).
      compute bootsz=mc.
    end if.
    loop.
      compute cilow = rnd(bootsz*(1-(conf/100))/2).
      compute cihigh = trunc((bootsz*(conf/100)+(bootsz*(1-(conf/100))/2)))+1.
      do if (cilow < 1 or cihigh > bootsz).
        compute bootsz=trunc((bootsz+1000)/1000)*1000.
        compute adjust = 1.
      end if.
    end loop if (cilow gt 0 and cihigh le bootsz).
    compute boot=bootsz.
    do if (mc > 0).
      compute mc = bootsz.
    end if.
    do if (boot > 0 and mc > 0 and (model > 3 and model < 6)).
      compute boot=0.
    end if.
    do if (adjust = 1 and boot > 0).
      compute note(notes,1)=6.
      compute notes = notes+1.
    end if.
    do if (adjust = 1 and mc > 0).
      compute note(notes,1)=13.
      compute notes = notes+1.
    end if.
  end if.
  do if (model = 6 and nmeds > 1).
    compute mmpaths = make((nmeds+2),(nmeds+2),0).
    do if (nmeds = 2).
      compute indboot = make(boot+1, 3, 999).
    else if (nmeds = 3).
      compute indboot = make(boot+1, 7, 999).
    else if (nmeds = 4).
      compute indboot = make(boot+1, 15, 999).
    end if.
    compute indlbl = {"Total:"; "Ind1 :"; "Ind2 :"; "Ind3 :"; "Ind4 :"; "Ind5 :"; "Ind6 :"; "Ind7 "+
    ":"; "Ind8 :"; "Ind9 :"; "Ind10:"; "Ind11:"; "Ind12:"; "Ind13:"; "Ind14:"; "Ind15:"}.
    compute indlbl2 = {"Ind1"; "Ind2"; "Ind3"; "Ind4"; "Ind5"; "Ind6"; "Ind7"; "Ind8"; "Ind9"; 
    "Ind10"; "Ind11"; "Ind12"; "Ind13"; "Ind14"; "Ind15"}.
    compute indces = make(boot+1, 4, 999).
  end if.
  do if (model < 4).
    compute boot = 0.
    compute cmat = make(10,1,0).
    compute zmat = make(10,1,0).
  end if.
  compute nvarch = make(1,ncol(dat),0).
  compute wmatch = 0.
  compute zmatch = 0.
  compute vmatch = 0.
  compute qmatch = 0.
  compute mmatch=0.
  compute minprobe=0.
  compute maxprobe=0.
  loop i = 1 to ncol(vnames).
    do if (vnames(:,i)=yname).
      compute y = dat(:,i).
      compute nvarch(1,i)=1.
      compute yerr = 0.
      do if ((yname = xname) or (yname = wname) or (yname = zname) or (yname = vname) or (yname = 
    qname)).
        compute errs = errs+1.
        compute runerrs(errs,1)=17.
      end if.
    end if.
    do if (vnames(:,i)=xname).
      compute x = dat(:,i).
      compute nvarch(1,i)=1.
      compute xdich = 1.
      compute xerr=0.
      loop jj = 1 to n.
         do if ((x(jj,1) <> cmin(x)) and (x(jj,1) <> cmax(x))).
           compute xdich = 0.
           break.
        end if.
      end loop.    
      compute xmean = csum(x)/n.
      do if (center = 1 and (model < 4 or xlist > 0) and mcx = 0).
        compute meanvec = make(n,1, xmean). 
        compute x = x-meanvec.
        compute centvar = {centvar, xname}.
      end if.
      compute xmean = csum(x)/n.
      compute tmp = x-(cons*xmean).
      compute xsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
      do if (xdich = 0).
         compute quantc(1,6) = 1.
        compute matx = {xmean-xsd; xmean; xmean+xsd}.
        do if (!xmodval = 999 and quantile = 0).
          do if ((matx(1,1) < cmin(x)) and model = 74).
             compute matx(1,1)=cmin(x).
             compute minprobe=1.    
           end if.
           do if ((matx(3,1) > cmax(x)) and model = 74).
             compute matx(3,1)=cmax(x).
             compute maxprobe=1.
           end if.
        end if.
        do if (quantile = 1).
          compute quantd(1,6) = 1.
          compute quantc(1,6) = 0.
          compute tmp = x.
          compute tmp(GRADE(x)) = x.
          compute matx = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
        end if.
      end if.
      do if (xdich = 1).
        compute matx = {cmin(x); cmax(x)}.
        do if (model=74).
          compute matx=cmin(x).
          compute mod74dic=1.
        end if.
      end if. 
      do if (!xmodval <> 999).
        compute matx = !xmodval.
        compute quantd(1,6) = 0.
        compute quantc(1,6) = 0.
      end if.
    end if.
    do if (werr = 0 and wlist = 1).
      do if (vnames(:,i)=wname).
        compute werr = 0.
        compute wmatch = 1.
        compute w = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(w)/n). 
          compute w = w-meanvec.
          compute centvar = {centvar, wname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute wmean = csum(w)/n.
        compute tmp = w-(cons*wmean).
        compute wsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute wdich = 1.
        loop jj = 1 to n.
          do if ((w(jj,1) <> cmin(w)) and (w(jj,1) <> cmax(w))).
            compute wdich = 0.
            break.
          end if.
        end loop. 
        do if (model = 3).
          compute jndich=wdich.
          compute jnmin=cmin(w).
          compute jnmax=cmax(w).
        end if.
        do if (wdich = 0).
          compute matw = {wmean-wsd; wmean; wmean+wsd}.
          do if (!wmodval = 999 and quantile = 0).
            do if (matw(1,1) < cmin(w)).
              compute matw(1,1)=cmin(w).
              compute minprobe=1.
            end if.
            do if (matw(3,1) > cmax(w)).
              compute matw(3,1)=cmax(w).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,1) = 1.
          do if (quantile = 1).
            compute quantd(1,1) = 1.
            compute quantc(1,1) = 0.
            compute tmp = w.
            compute tmp(GRADE(w)) = w.
            compute matw = {tmp(trunc(n*.10),1);tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (wdich = 1).
          compute matw = {cmin(w); cmax(w)}.
          compute wvdich=1.
          compute cmaxw=cmax(w).
          compute cminw=cmin(w).
        end if. 
        do if (!wmodval <> 999).
          compute matw = !wmodval.
          compute quantd(1,1) = 0.
          compute quantc(1,1) = 0.
        end if.
        compute modmatv(1,1)=nrow(matw).
        compute modmat((1:nrow(matw)), 1) = matw.
        compute modvnm(1,1)=wname.
        compute modmatp(1,1) = 1.
      end if.
    end if.
    do if (zerr = 0 and zlist = 1).
      do if (vnames(:,i)=zname).
        compute zerr = 0.
        compute zmatch = 1.
        compute z = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(z)/n). 
          compute z = z-meanvec.
          compute centvar = {centvar, zname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute zmean = csum(z)/n.
        compute tmp = z-(cons*zmean).
        compute zsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute zdich = 1.
        loop jj = 1 to n.
          do if ((z(jj,1) <> cmin(z)) and (z(jj,1) <> cmax(z))).
            compute zdich = 0.
            break.
          end if.
        end loop. 
        do if (zdich = 0).
          compute matz = {zmean-zsd; zmean; zmean+zsd}.
          do if (!zmodval = 999 and quantile = 0).
            do if (matz(1,1) < cmin(z)).
              compute matz(1,1)=cmin(z).
              compute minprobe=1.
            end if.
            do if (matz(3,1) > cmax(z)).
              compute matz(3,1)=cmax(z).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,2) = 1.
          do if (quantile = 1).
            compute quantd(1,2) = 1.
            compute quantc(1,2) = 0.
            compute tmp = z.
            compute tmp(GRADE(z)) = z.
            compute matz = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (zdich = 1).
          compute matz = {cmin(z); cmax(z)}.
        end if. 
        do if (!zmodval <> 999).
          compute matz = !zmodval.
          compute quantd(1,2) = 0.
          compute quantc(1,2) = 0.
        end if.
        compute modmatv(1,2)=nrow(matz).
        compute modmat((1:nrow(matz)), 2) = matz.
        compute modvnm(1,2)=zname.
        compute modmatp(1,2) = 1.
      end if.
    end if.
    do if (verr = 0 and vlist = 1).
      do if (vnames(:,i)=vname).
        compute verr = 0.
        compute vmatch = 1.
        compute v = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(v)/n). 
          compute v = v-meanvec.
          compute centvar = {centvar, vname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute vmean = csum(v)/n.
        compute tmp = v-(cons*vmean).
        compute vsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute vdich = 1.
        loop jj = 1 to n.
          do if ((v(jj,1) <> cmin(v)) and (v(jj,1) <> cmax(v))).
            compute vdich = 0.
            break.
          end if.
        end loop.
        do if (vdich = 0).
          compute matv = {vmean-vsd; vmean; vmean+vsd}.
          do if (!vmodval = 999 and quantile = 0).
            do if (matv(1,1) < cmin(v)).
              compute matv(1,1)=cmin(v).
              compute minprobe=1.
            end if.
            do if (matv(3,1) > cmax(v)).
              compute matv(3,1)=cmax(v).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,3) = 1.
          do if (quantile = 1).
            compute quantd(1,3) = 1.
            compute quantc(1,3) = 0.
            compute tmp = v.
            compute tmp(GRADE(v)) = v.
            compute matv = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (vdich = 1).
          compute matv = {cmin(v); cmax(v)}.
          compute wvdich=1.
          compute cmaxv=cmax(v).
          compute cminv=cmin(v).
        end if. 
        do if (!vmodval <> 999).
          compute matv = !vmodval.
          compute quantd(1,3) = 0.
          compute quantc(1,3) = 0.
        end if.
        compute modmatv(1,3)=nrow(matv).
        compute modmat((1:nrow(matv)), 3) = matv.
        compute modvnm(1,3)=vname.
        compute modmatp(1,3) = 1.
      end if.
    end if.
    do if (qerr = 0 and qlist = 1).
      do if (vnames(:,i)=qname).
        compute qerr = 0.
        compute qmatch = 1.
        compute q = dat(:,i).
        do if (center = 1).
          compute meanvec = make(n,1,csum(q)/n). 
          compute q = q-meanvec.
          compute centvar = {centvar, qname}.
        end if.
        compute nvarch(1,i)=1.
        compute nmods = nmods + 1.
        compute qmean = csum(q)/n.
        compute tmp = q-(cons*qmean).
        compute qsd = sqrt((1/(n-1))*(t(tmp)*tmp)).
        compute qdich = 1.
        loop jj = 1 to n.
          do if ((q(jj,1) <> cmin(q)) and (q(jj,1) <> cmax(q))).
            compute qdich = 0.
            break.
          end if.
        end loop.
        do if (qdich = 0).
          compute matq = {qmean-qsd; qmean; qmean+qsd}.
          do if (!qmodval = 999 and quantile = 0).
            do if (matq(1,1) < cmin(q)).
              compute matq(1,1)=cmin(q).
              compute minprobe=1.
            end if.
            do if (matq(3,1) > cmax(q)).
              compute matq(3,1)=cmax(q).
              compute maxprobe=1.
            end if.
          end if.
          compute quantc(1,4) = 1.
          do if (quantile = 1).
            compute quantd(1,4) = 1.
            compute quantc(1,4) = 0.
            compute tmp = q.
            compute tmp(GRADE(q)) = q.
            compute matq = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
          end if.
        end if.
        do if (qdich = 1).
          compute matq = {cmin(q); cmax(q)}.
        end if. 
        do if (!qmodval <> 999).
          compute matq = !qmodval.
          compute quantd(1,4) = 0.
          compute quantc(1,4) = 0.
        end if.
        compute modmatv(1,4)=nrow(matq).
        compute modmat((1:nrow(matq)), 4) = matq.
        compute modvnm(1,4)=qname.
        compute modmatp(1,4) = 1.
      end if.
    end if.
    do if (vnames(:,i)=clname).
      compute cld = dat(:,i).
      compute cvname = vnames(:,i).
      compute nvarch(1,i)=1.
      compute clsmtch = 1.
    end if.
    loop j = 1 to ncol(mnames).
      do if (vnames(:,i)=mnames(1,j)).
        compute mmatch = mmatch + 1.
        compute m(:,j)=dat(:,i).
        do if (center = 1 and nvarch(1,i) = 0 and mlist > 0 and mcm = 0).
          compute meanvec = make(n,1,csum(m(:,j))/n). 
          compute m(:,j) = m(:,j)-meanvec.
          compute centvar = {centvar, mnames(1,j)}.
        end if.
        compute nvarch(1,i)=1.
        compute dichm = 1.
        loop jj = 1 to n.
          do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
            compute dichm = 0.
            break.
          end if.
        end loop.
        do if ((dichm = 1) and (olsdichm = 1)).
          compute note(notes,1) = 18.
          compute notes = notes + 1.
        end if.
        do if (dichm = 1 and model > 3 and mcheck = 0 and olsdichm = 0).
          compute errs= errs+1.
          compute runerrs(errs,1) = 1.
          compute mcheck = 1.
        end if.
        do if ((model <= 3) and (ncol(mnames)=1)).
          compute nmods = nmods + 1.
          compute mmean = csum(m(:,j))/n.
          compute tmp = m(:,j)-(cons*mmean).
          compute msd = sqrt((1/(n-1))*(t(tmp)*tmp)).
          compute mdich = 1.
          loop jj = 1 to n.
            do if ((m(jj,j) <> cmin(m(:,j))) and (m(jj,j) <> cmax(m(:,j)))).
              compute mdich = 0.
              break.
            end if.
          end loop.
          do if (model = 1).
            compute jndich=mdich.
            compute jnmin=cmin(m(:,j)).
            compute jnmax=cmax(m(:,j)).
          end if.
          do if (mdich = 0).
            compute matm = {mmean-msd; mmean; mmean+msd}.
            do if (!mmodval = 999 and quantile = 0).
              do if (matm(1,1) < cmin(m(:,j))).
                compute matm(1,1)=cmin(m(:,j)).
                compute minprobe=1.
              end if.
              do if (matm(3,1) > cmax(m(:,j))).
                compute matm(3,1)=cmax(m(:,j)).
                compute maxprobe=1.
              end if.
            end if.
            compute quantc(1,5) = 1.
            do if (quantile = 1).
              compute quantd(1,5) = 1.
              compute quantc(1,5) = 0.
              compute tmp = m(:,j).
              compute tmp(GRADE(m(:,j))) = m(:,j).
              compute matm = {tmp(trunc(n*.10),1); tmp(trunc(n*.25),1); tmp(trunc(n*.50),1); 
    tmp(trunc(n*0.75),1); tmp(trunc(n*.90),1)}.
            end if.
          end if.
          do if (mdich = 1).
            compute matm = {cmin(m); cmax(m)}.
          end if. 
          do if (!mmodval <> 999).
            compute matm = !mmodval.
            compute quantd(1,5) = 0.
            compute quantc(1,5) = 0.
          end if.
          compute modmatv(1,5)=nrow(matm).
          compute modmat((1:nrow(matm)), 5) = matm.
          compute modvnm(1,5)=mnames(1,j).
          compute modmatp(1,5) = 1.
        end if.
      end if.
    end loop.
  end loop.
  do if (minprobe=1).
    compute note(notes,1) = 14.
    compute notes = notes + 1.
  end if.
  do if (maxprobe=1).
    compute note(notes,1) = 15.
    compute notes = notes + 1.
  end if.
  do if (clname <> "xxx" and clsmtch = 0).
    compute errs = errs+1.
    compute runerrs(errs,1) = 23.
  end if.
  do if (clname <> "xxx").
    do if ((clname = zname) or (clname = vname) or (clname = wname) or (clname = xname) or (clname 
    = yname) or (clname = qname)).
      compute errs = errs+1.
      compute runerrs(errs,1) = 24.
    end if.
  end if.
  do if (wlist = 1 and werr = 0 and wmatch = 0).
    compute werr = 3.
     compute errs = errs+1.
     compute runerrs(errs,1) = 4.
  end if.
  do if (zlist = 1 and zerr = 0 and zmatch = 0).
    compute zerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 5.
  end if.
  do if (qlist = 1 and qerr = 0 and qmatch = 0).
    compute qerr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 6.
  end if.
  do if (vlist = 1 and verr = 0 and vmatch = 0).
    compute verr = 3.
    compute errs = errs+1.
    compute runerrs(errs,1) = 7.
  end if.
  do if (yerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 16.
  end if.
  do if (xerr = 1).
    compute errs = errs+1.
    compute runerrs(errs,1) = 32.
  end if.
  do if (model = 6 and nmeds < 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 18.
  end if.
  do if (mmatch < ncol(mnames)).
    compute errs=errs+1.
    compute runerrs(errs,1) = 25.
  end if.
  do if (clname <> "xxx").
    compute cld = design(cld).
    compute cluster = ncol(cld).
    compute cld = cld(:,2:ncol(cld)).
    compute clsdmy = ncol(cld).
    do if (clsdmy > 19).
      compute errs = errs+1.
      compute runerrs(errs,1) = 26.
    end if.
    do if (mcx > 0 or mcm > 0).
      compute errs = errs+1.
      compute runerrs(errs,1) = 38.
    end if.
  end if.
  compute dichy = 1.
  loop jj = 1 to n.
    do if ((y(jj,1) <> cmin(y)) and (y(jj,1) <> cmax(y))).
       compute dichy = 0.
       break.
    end if.
  end loop.    
  do if ((dichy = 1) and (olsdichy = 1)).
    compute dichy=0.
    compute note(notes,1) = 18.
    compute notes = notes + 1.
  end if.
  do if (dichy = 1).
    compute jncrit=xp2*xp2.
    do if (ws = 1).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 28.
    end if.
  end if.
  compute ncovs = ncol(dat)-rsum(nvarch).
  do if (effsize = 1 and covmy <> 0 and ncovs > 0 and model > 3 and model < 7).
    compute note(notes,1) = 22.
    compute notes = notes + 1.
    compute effsize = 0.
  end if.
  do if (ws = 1 and ncovs < nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 29.
  end if. 
  do if (ws = 1 and ncovs > nmeds).
    compute criterr = 1.
    compute errs = errs+1.
    compute runerrs(errs,1) = 30.
  end if. 
end if.
do if (errs = 0).
  do if (rsum(quantd) > 0).
    compute note(notes,1) = 4.
    compute notes = notes+1.
  end if.
  do if (rsum(quantc) > 0).
    compute note(notes,1) = 5.
    compute notes = notes+1.
  end if.
  do if (ncovs > 0).
    compute c = make(n,ncovs,0).
    compute cnames = {"x"}.
    compute j = 1.
    loop i = 1 to ncol(vnames).
      do if (nvarch(1,i)) = 0.
        compute c(:,j) = dat(:,i).
        compute nvarch(1,i)=1.
        compute j=j+1.
        compute cnames = {cnames, vnames(:,i)}.
      end if.
    end loop.
    compute cnames = cnames(1,2:ncol(cnames)).
    do if (ws = 1).
      compute covmean=csum(c)/n.
      loop i = 1 to ncovs.
        compute meanvec = make(n,1, covmean(1,i)). 
        compute c(:,i) = c(:,i)-meanvec.
      end loop.
      compute centvar = {centvar, cnames}.
      compute x=c(:,1:ncovs).
      do if (ncovs=nmeds).
      compute ncovs=0.
      end if.
    end if.
  end if.
  compute names = {yname, xname, mnames, wname, zname, vname, qname}.
  do if (ncovs > 0).
    compute names = {names, cnames}.
  end if.
  do if (dichy = 1 and effsize = 1).
    compute note(notes,1) = 2.
    compute notes = notes+1.
  end if.
  do if (model > 3 and model < 6). 
    compute indeff=make(nmeds,1,0).
    compute indboot=make(boot+1,nmeds,999).
    do if (mc > 0).
      compute indboot=make(mc+1,nmeds,999).      
    end if.
    do if (effsize = 1 and dichy = 0).
      compute rmeff=make(boot+1,nmeds+1,999).
      compute abpseff=make(boot+1,nmeds+1,999).
      compute abcseff=make(boot+1,nmeds+1,999).
      compute pmeff=make(boot+1,nmeds+1,999).
      compute r245 = make(boot+1,1,999).
      compute kappa2 = make(boot+1,1,999).
    end if.
  end if.
  do if (model = 6 and effsize = 1 and dichy = 0).
    compute rmeff=make(boot+1,ncol(indboot),999).
    compute abpseff=make(boot+1,ncol(indboot),999).
    compute abcseff=make(boot+1,ncol(indboot),999).
    compute pmeff=make(boot+1,ncol(indboot),999).
  end if.
  do if (nmods > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm(1,tmp) = modvnm(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:nmods).
    compute modvnm=modvnm(:,1:nmods).
    compute modmatv=modmatv(:,1:nmods).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvals = make((modmatv(1,1)*modprod(1,1)), nmods,0).
    loop i = 1 to nmods.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvals)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvals(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
    do if (model = 74).
      compute modvals=matx.
      compute modvnm=xname.
    end if.
    compute vmat = make(8,nrow(modvals),0).
    compute vmat(1,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute vmat(5,1:nrow(modvals)) = make(1,nrow(modvals),1).
    compute indeff=make(nrow(modvals),1,0).
    do if (model <> 5).
      compute indboot = make(((boot+1)*nmeds), nrow(modvals), -99999999).
      compute indbootp = make(boot+1,nmeds, -99999999).
    end if.
  end if.
  do if (nmods> 0).
    loop i = 1 to ncol(modvals).
      do if (modvnm(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
  end if.
  do if (dichy = 1).
    compute omx = cmax(y).
    compute omn = cmin(y).
    compute y = (y = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute data = {cons,y,m,x}.
  compute datamed = data.
  compute datayed = data.
  compute datanm = {"constant"; yname; t(mnames); xname}.
  compute datanmm = {"constant"; yname; t(mnames); xname}.
  compute datanmy = {"constant"; yname; t(mnames); xname}.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  do if (model < 4 and errs = 0).
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, " ", " "}.
    compute datayed = {datayed, x&*m}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(1,i) = 1.
      compute vmat(2,i) = modvals(i,1).
    end loop.
    compute mmat = make(16,nmeds,1).
  end if.
  do if (model = 2 or model = 3).
    compute int1 = x&*w.
    compute datayed = {datayed, w, int1}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,2).
      compute vmat(3,i) = modvals(i,1).
      compute vmat(4,i) = modvals(i,1)*modvals(i,2).
    end loop.
  end if.
  do if (model = 3).
    compute yintkey = {yintkey; yintemp(1,intcnt), mnames, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute int1 = w&*m.
    compute int2 = x&*w&*m.
    compute datayed = {datayed, int1, int2}.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", mnames, "   X", wname}. 
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (model = 4 or model = 5).
    compute vmat = make(8,1,1).
  end if.
  compute yintkey2 = yintkey.
  do if (wm = 1).
    compute int1 = x&*w.
    compute datamed = {datamed, w,int1}.
    compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmm = {datanmm; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    loop i = 1 to nrow(modvals).
      compute vmat(2,i) = modvals(i,wcol).
    end loop.
    do if (zm = 1).
      compute int1 = x&*z.
      compute datamed = {datamed, z, int1}.
      compute yintkey = {yintkey;  yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; zname; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      loop i = 1 to nrow(modvals).
        compute vmat(3,i) = modvals(i,zcol).
      end loop.
    end if.  
    do if (wzm = 1).
      compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
      compute datanmm = {datanmm; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      compute int1 = w&*z.
      compute int2 = x&*w&*z.
      compute datamed = {datamed, int1, int2}.
      loop i = 1 to nrow(modvals).
        compute vmat(4,i) = modvals(i,wcol)*modvals(i,zcol).
      end loop.
    end if.
  end if.
  compute mdatacol=ncol(datamed).
  compute mintkey = yintkey.
  compute yintkey = {" ", " ", " ", " ", " ", " "}.
  compute medints = intcnt-1.
  do if (vy = 1 or xmy = 1).
    compute mp = 1.
    do if (xmy = 1).
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,1).
      end loop.
    end if.
    do if (vy = 1).
      compute datayed = {datayed, v}.
      compute datanmy = {datanmy; vname}.
      compute mmods = 1.
      loop i = 1 to nrow(modvals).
        compute vmat(6,i) = modvals(i,vcol).
      end loop.
      do if (qy = 1).
        compute mp = 2.
        compute datayed = {datayed,q}.
        compute datanmy = {datanmy; qname}.
        compute mmods = 2.
        loop i = 1 to nrow(modvals).
          compute vmat(7,i) = modvals(i,qcol).
        end loop.
      end if.
      do if (vqy = 1).
        compute mp = 3.
        compute datayed = {datayed, v&*q}.
        compute mmods = 3.
        loop i = 1 to nrow(modvals).
          compute vmat(8,i) = modvals(i,vcol)*modvals(i,qcol).
        end loop.
      end if.
    end if.
    compute mints = make(n,(nmeds*mp),0).
    loop i = 0 to (nmeds-1).
      do if (i = 0 and vqy = 1).
        compute yintkey = {yintkey; yintemp(1,intcnt), vname, "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (vy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (xmy = 1).
        compute mints(:,((i*mp)+1))= m(:,(i+1))&*x.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", xname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      do if (qy = 1).
        compute mints(:,((i*mp)+2))=m(:,(i+1))&*q.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", qname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
        do if (vqy = 1).
          compute mints(:,((i*mp)+3))=m(:,(i+1))&*v&*q.
          compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", vname, "   X", 
    qname}.
          compute datanmy = {datanmy; yintemp(1,intcnt)}.
          compute intcnt = intcnt+1.
        end if.
      end if.            
    end loop.
    compute datayed = {datayed, mints}.
  end if.
  compute mp = 1.
  do if (wvmy = 1).
    compute mp = 2.
    loop i = 1 to nrow(modvals).
      compute vmat(8,i) = modvals(i,wcol)*modvals(i,vcol).
    end loop.
  end if.
  compute mints2 = make(n,(nmeds*mp),0).
  do if (wmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(7,i) = modvals(i,wcol).
    end loop.
    do if (wy = 0 and model > 3).
      compute datayed = {datayed, w}.
      compute datanmy = {datanmy; wname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wvmy = 1).
        compute datayed = {datayed,w&*v}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", vname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints2(:,((i*mp)+1))= m(:,(i+1))&*w.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wvmy = 1).
        compute mints2(:,((i*mp)+2))= m(:,(i+1))&*w&*v.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    vname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints2}.
  end if.
  compute mp = 1.
  do if (zmy = 1).
    loop i = 1 to nrow(modvals).
      compute vmat(6,i) = modvals(i,zcol).
    end loop.
    do if (wzmy = 1).
      compute mp = 2.
      loop i = 1 to nrow(modvals).
        compute vmat(8,i) = modvals(i,zcol)&*modvals(i,wcol).
      end loop.
    end if.
  end if.
  do if (zmy = 1).
    compute mints3 = make(n,(nmeds*mp),0).
    do if (zy = 0).
      compute datayed = {datayed, z}.
      compute datanmy = {datanmy; zname}.
    end if.
    loop i = 0 to (nmeds-1).
      do if (i = 0 and wzmy = 1 and wzy = 0).
        compute datayed = {datayed,w&*z}.
        compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
      compute mints3(:,((i*mp)+1))= m(:,(i+1))&*z.
      compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", zname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (wzmy = 1).
        compute mints3(:,((i*mp)+2))= m(:,(i+1))&*w&*z.
        compute yintkey = {yintkey; yintemp(1,intcnt), mnames(1,(i+1)), "   X", wname, "   X", 
    zname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end loop.
    compute datayed = {datayed, mints3}.
  end if.
  compute decoc = 1.
  compute modmat = make(5,5,999).
  compute modmatv = make(1,5,1).
  compute modmatp = make(1,5,0).
  compute modprod = modmatv.
  do if (wy = 1 and model > 3).
    compute datayed = {datayed, w, x&*w}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,1)=nrow(matw).
    compute modmat((1:nrow(matw)), 1) = matw.
    compute modvnm2(1,1)=wname.
    compute modmatp(1,1) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, " ", " "}.
    compute datanmy = {datanmy; wname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (zy = 1).
    compute datayed = {datayed,z,x&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,2)=nrow(matz).
    compute modmat((1:nrow(matz)), 2) = matz.
    compute modvnm2(1,2)=zname.
    compute modmatp(1,2) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy; zname; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (wzy = 1).
    compute datayed = {datayed,w&*z,x&*w&*z}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), wname, "   X", zname, " ", " "}.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "   X", zname }.
    compute datanmy = {datanmy;yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  do if (vxy = 1).
    compute datayed = {datayed, x&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute modmatv(1,3)=nrow(matv).
    compute modmat((1:nrow(matv)), 3) = matv.
    compute modvnm2(1,3)=vname.
    compute modmatp(1,3) = 1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, " ", " "}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
    do if (qxy = 1).
      compute datayed = {datayed, x&*q}.
      compute decoc = decoc+1.
      compute deco(decoc,1) = ncol(datayed)-1.
      compute modmatv(1,4)=nrow(matq).
      compute modmat((1:nrow(matq)), 4) = matq.
      compute modvnm2(1,4)=qname.
      compute modmatp(1,4) = 1.
      compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", qname, " ", " "}.
      compute datanmy = {datanmy; yintemp(1,intcnt)}.
      compute intcnt = intcnt+1.
      do if (vqxy = 1).
        compute datayed = {datayed, x&*v&*q}.
        compute decoc = decoc+1.
        compute deco(decoc,1) = ncol(datayed)-1.
        compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", vname, "X", qname}.
        compute datanmy = {datanmy; yintemp(1,intcnt)}.
        compute intcnt = intcnt+1.
      end if.
    end if.
  end if.
  do if (wvxy = 1).
    compute datayed = {datayed, x&*w&*v}.
    compute decoc = decoc+1.
    compute deco(decoc,1) = ncol(datayed)-1.
    compute yintkey = {yintkey; yintemp(1,intcnt), xname, "   X", wname, "X", vname}.
    compute datanmy = {datanmy; yintemp(1,intcnt)}.
    compute intcnt = intcnt+1.
  end if.
  compute modvalsd = 0.
  compute ttt = rsum(modmatp).
  do if (rsum(modmatp) > 0).
    compute tmp = 1.
    loop i = 1 to 5.
      do if (modmatp(1,i) = 1).
        compute modmat(:,tmp) = modmat(:,i).
        compute modvnm2(1,tmp) = modvnm2(1,i).
        compute modmatv(1,tmp) = modmatv(1,i).
        compute tmp=tmp+1.
      end if.
    end loop.
    compute modmat=modmat(:,1:ttt).
    compute modvnm2=modvnm2(:,1:ttt).
    compute modmatv=modmatv(:,1:ttt).
    loop i = 1 to (ncol(modmatv)-1).
      compute tmp = 1.
      loop j = (i+1) to ncol(modmatv).
        compute tmp = tmp*modmatv(1,j).
      end loop.
      compute modprod(1,i)=tmp.
    end loop.
    compute modvalsd = make((modmatv(1,1)*modprod(1,1)), ttt,0).
    loop i = 1 to ttt.
      compute strt = 1.
      compute fnsh=0.
      loop if (fnsh < nrow(modvalsd)).
        loop j = 1 to modmatv(1,i).
          compute tmp=make(modprod(1,i),1,modmat(j,i)).
          compute fnsh = fnsh+nrow(tmp).
          compute modvalsd(strt:fnsh, i) = tmp.
          compute strt = fnsh+1.
        end loop.
      end loop.
    end loop.
  end if.
  do if (ttt > 0).
    loop i = 1 to ncol(modvalsd).
      do if (modvnm2(1,i)=wname).
        compute wcol = i.
      end if.
      do if (modvnm2(1,i)=zname).
        compute zcol = i.
      end if.
      do if (modvnm2(1,i)=vname).
        compute vcol = i.
      end if.
      do if (modvnm2(1,i)=qname).
        compute qcol = i.
      end if.
    end loop.
    compute directv = make(nrow(modvalsd),1,1).
    do if (wy = 1).
      compute directv ={directv, modvalsd(:,wcol)}.
    end if.
    do if (zy = 1).
      compute directv ={directv, modvalsd(:,zcol)}.
    end if.
    do if (wzy = 1).
      compute directv ={directv, (modvalsd(:,wcol)&*modvalsd(:,zcol))}.
    end if.
    do if (vxy = 1).
      compute directv ={directv, modvalsd(:,vcol)}.
    end if.
    do if (qxy = 1).
      compute directv ={directv, modvalsd(:,qcol)}.
    end if.
    do if (vqxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,qcol))}.
    end if.
    do if (wvxy = 1).
      compute directv ={directv, (modvalsd(:,vcol)&*modvalsd(:,wcol))}.
    end if.
  end if.
  compute ydatacol=ncol(datayed).
  do if (ncovs > 0).
    do if (covmy <> 2).
      compute datamed = {datamed,c}.
    end if.
    do if (covmy <> 1).
      compute datayed = {datayed,c}.
    end if.
     /* begin added */.
    do if (covmy = 1 and model = 4 and mcx > 0).
      compute datayed = {datayed,c}.
    end if.
      /* end added */.
    compute covmeans = csum(c)/n.
  end if.
  do if (cluster > 0).
    compute datamed = {datamed,cld}.
    compute datayed = {datayed, cld}.
    compute cldmeans = csum(cld)/n.
  end if.
  compute mst = 3.
  compute mnd = mst+nmeds-1.
  compute ydatacol=ncol(datayed).
  compute mdatacol=ncol(datamed).
  do if (ncovs > 0).
    compute datanmy = {datanmy; t(cnames)}.
    do if (model > 3).
      compute datanmm = {datanmm; t(cnames)}.
    end if.
  end if.
  compute datanmy = {"constant"; datanmy(3:nrow(datanmy),1)}.
  do if (model > 3).
    compute datanmm = {"constant"; datanmm(3:nrow(datanmm),1)}.
  end if.
  compute amm = make(2,1,0).
  compute abmm = make(2,1,0).
  compute mnv = csum(datayed(:,2)/n).
  compute mnv = make(n,1,mnv).
  compute ssty = csum((datayed(:,2)-mnv)&**2).
  compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
  compute sigma = sigma/(n*(n-1)).
  do if (ws=1).
    compute stddevm=diag(sigma).
    compute stddevm=sqrt(stddevm(3:(nmeds+2),1)).
  end if.
  compute stddevy = sqrt(sigma(2,2)).
  compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
  compute r2xy = (sigma(2,(3+nmeds))/(stddevy*stddevx))&**2.
  compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
  compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
  do if (model = 4 and nmeds = 1 and cluster = 0 and ncovs = 0).
    compute kappaa = sigma(2,3)*sigma(2,4).
    compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
    compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
    compute kappad = sigma(4,4)*sigma(2,2).
    compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
    compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
    compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
    do if (sigma(3,4) < 0).
      compute amma =cmin(amm).
    end if.
    do if (sigma(3,4) > 0).
      compute amma = cmax(amm).
    end if.
    compute abmm(1,1)=-amma*(kappac/kappae).
    compute abmm(2,1)=amma*(kappac/kappae).
  end if.
  compute datatm = datamed.
  compute dataty = datayed.
  compute mdlnms2 = {!quote(!model); yname; xname}.
  compute mdlnms = {"Model ="; "    Y ="; "    X =" }.
  do if (ws = 1).
  compute mdlnms2 = {!quote(!model); yname}.
  compute mdlnms = {"Model ="; "    Y ="}.
  end if.
  loop i = 1 to ncol(mnames).
    compute mdlnms2 = {mdlnms2; mnames(1,i)}.
    do if (i = 1 and ncol(mnames) = 1).
      compute mdlnms = {mdlnms; "    M ="}.
    else.
      compute mdlnms = {mdlnms; mlab(i,1)}.
    end if.
  end loop.
  do if (wname <> "xxx").
    compute mdlnms2 = {mdlnms2; wname}.
    compute mdlnms = {mdlnms; "    W = "}.
  end if.
  do if (zname <> "xxx").
    compute mdlnms2 = {mdlnms2; zname}.
    compute mdlnms = {mdlnms; "    Z = "}.
  end if.
  do if (vname <> "xxx").
    compute mdlnms2 = {mdlnms2; vname}.
    compute mdlnms = {mdlnms; "    V = "}.
  end if.
  do if (qname <> "xxx").
    compute mdlnms2 = {mdlnms2; qname}.
    compute mdlnms = {mdlnms; "    Q = "}.
  end if.
  do if (jn = 1 and model = 1 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  do if (jn = 1 and model = 3 and jndich = 1).
    compute note(notes,1) = 8.
    compute notes = notes + 1.
  end if.
  compute mdpbe=0.
  compute medte2=0.
  do if (!model=4 and mcx > 0).
    compute medte2=1.
  end if.  
  do if (((mcx > 0) or (mcm > 0)) AND model = 1).
     compute mdpbe=1.
     compute dd={datayed(:,4),datayed(:,3)}.
     do if (ncovs > 0).
       compute dd={datayed(:,6:(5+ncovs)),dd}.
     end if.
     compute dd={datayed(:,2),dd}.
  end if.
  compute yes = ((nrow(modvals)=1) and (contrast = 1) and (model = 3 or model = 2)).
  compute yes2 = ((nrow(modvals)=1) and (contrast = 1) and (model = 2)).
  print/title = "************** PROCESS Procedure for SPSS Release 2.15 *******************".
  print/title = "          Written by Andrew F. Hayes, Ph.D.       www.afhayes.com".
  print/title = "    Documentation available in Hayes (2013). www.guilford.com/p/hayes3"/space=0.
  print mdlnms2/title = "**************************************************************************"+
    ""/rnames = mdlnms/format = a8.
  do if (ncovs > 0).
    print cnames/title = "Statistical Controls:"/rlabels = "CONTROL="/format a8.
  end if.
  print n/title = "Sample size"/format F10.0.
  do if (cluster > 0).
    print cluster/rnames = cvname/title = "Clustering Variable and Number of Clusters".
  end if.
  do if (model > 3 and (!quote(!seed) <> "random")).
    compute seedt=!quote(!seed).
    print seedt/title="Custom"/format=A12/rlabel = "Seed:".
  end if.
  do if (mdpbe=0 and medte2 = 0).
    loop bt = 1 to (boot+1).
       do if (bt = 2 and savboot = 1).
          compute bootstrp = make(boot,(ncol(bootcoef)),-999).
       end if.
       compute bootcoef = 0.  
      do if (bt > 1).
        loop.
          compute v=trunc(uniform(n,1)*n)+1.
          compute datayed = dataty(v,:).
          do if (ws=1).
             compute bmeans=csum(datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed)))/n.
             compute cntmat=make(n,nmeds,1).
             compute bmeans2=bmeans.
             do if (nmeds > 1).
                 compute bmeans2=mdiag(bmeans).
             end if.
             compute cntmat=cntmat*bmeans2.
             compute datayed(:,(ncol(datayed)-nmeds+1):ncol(datayed))=datayed(:,(ncol(datayed)
    -nmeds+1):ncol(datayed))-cntmat.
          end if.
          compute detcheck=det(t(datayed)*datayed).
          compute rk=(detcheck <> 0).
          do if (model > 3).
            compute datamed = datatm(v,:). 
            compute detcheck=det(t(datamed)*datamed).
            do if (rk=1).
              compute rk=(detcheck <> 0).
            end if.
          end if.
          compute sigma = (n*sscp(datayed))-(t(csum(datayed))*(csum(datayed))).
          compute sigma = sigma/(n*(n-1)).
          compute temp=diag(sigma).
          compute bad = bad+(1-rk).
          compute false = 1.
        end loop if (rk = 1).
        compute stddevy = sqrt(sigma(2,2)).
        compute stddevx = sqrt(sigma((3+nmeds),(3+nmeds))).
        compute ctot = sigma(2,(3+nmeds))/sigma((3+nmeds),(3+nmeds)).
        do if (model = 4 and nmeds = 1 and ncovs = 0 and cluster = 0).
          compute r2xy = (sigma(2,4)/(stddevy*stddevx))&**2.
          compute r2my = (sigma(2,3)/(stddevy*sqrt(sigma(3,3))))&**2.
          compute sstot = sigma(2,2)*(n-1).
          compute kappaa = sigma(2,3)*sigma(2,4).
          compute kappab = sqrt((sigma(3,3)*sigma(2,2))-(sigma(2,3)*sigma(2,3))).
          compute kappac = sqrt((sigma(4,4)*sigma(2,2))-(sigma(2,4)*sigma(2,4))).
          compute kappad = sigma(4,4)*sigma(2,2).
          compute kappae = sqrt((sigma(4,4)*sigma(3,3))-(sigma(3,4)*sigma(3,4))).
          compute amm(1,1) = (kappaa+(kappab*kappac))/kappad.
          compute amm(2,1) = (kappaa-(kappab*kappac))/kappad.
          do if (sigma(3,4) < 0).
            compute amma =cmin(amm).
          end if.
          do if (sigma(3,4) > 0).
            compute amma = cmax(amm).
          end if.
          compute abmm(1,1)=-amma*(kappac/kappae).
          compute abmm(2,1)=amma*(kappac/kappae).
        end if.
      end if.
      /* estimate model of mediator(s) */.
      do if (model > 3).
        loop im = 1 to nmeds.
          compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
          compute xmnm = {"constant"; datanmm((2+nmeds):nrow(datanmm),1)}.
          compute invXtX = inv(t(xm)*xm).
          compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
          do if (model = 6).
            do if (im = 1).
              compute xm={cons, datamed(:,(mnd+1):mdatacol)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
            end if.
            do if (im > 1).
              compute xm={cons, datamed(:,3:(im+1)), datamed(:,(mnd+1):mdatacol)}.
              compute xmnm = {"constant"; datanmm(2:im,1); datanmm((mnd):nrow(datanmm),1)}.
              compute invXtX = inv(t(xm)*xm).
              compute coeff =invXtX*t(xm)*datamed(:,(2+im)).
              compute mmpaths((im+1),(2:im))=t(coeff(2:im,1)).
            end if.
          end if.
          do if (ws = 1).
          compute coeff(1,1)=csum(datamed(:,(2+im)))/n.
          end if.
          compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          do if (bt = 1).
            compute resid=datamed(:,(2+im))-xm*coeff.
            compute sse =cssq(resid).
            compute mse = sse/(n-ncol(xm)).
            compute mnv = csum(data(:,(2+im))/n).
            compute mnv = make(n,1,mnv).
            compute sstm = csum((data(:,(2+im))-mnv)&**2).
            compute k3 = nrow(coeff).
            do if (hc3 = 1).
              compute h = xm(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xm(i3,:)*invXtX*t(xm(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xm(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xm(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xm(:,i3) = sqrt(mse)&*xm(:,i3).
              end loop.
            end if.
            compute lmat = ident(nrow(coeff)).
            compute lmat = lmat(:,2:ncol(lmat)).
            compute hccov=invXtX*t(xm)*xm*invXtX.
            compute mcmats(im,im)=hccov(2,2).
            compute dfnum = nrow(coeff)-1.
            compute dfden = n-dfnum-1.
            compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*hccov*lmat)*((t(lmat)*coeff)))/dfnum.
            compute coeff = coeff(1:(nrow(coeff)-clsdmy),1).
            compute mccoeff(im,1)=coeff(2,1).
            compute standerr = sqrt(diag(invXtX*t(xm)*xm*invXtX)).
            compute standerr = standerr(1:(nrow(standerr)-clsdmy),1).
            do if (ws=1).
              compute standerr(1,1)=stddevm(im,1)/sqrt(n).
            end if.
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xm)))).
            compute temp=(n-ncol(xm)).
            do if (ws=1).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute temp=(n-1).
            end if.
            compute xd = abs(xp2).
            compute temp =  (temp* (exp((temp-(5/6))*((xd/(temp-(2/3)+(.11/temp)))*(xd/(temp-(2/3)+
    (.11/temp)))))-1)).
            compute temp1 = coeff-sqrt(abs(temp))*standerr.
            compute temp2 = coeff+sqrt(abs(temp))*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (ws=1).
              compute op=op(1,:).
              compute xmnm = apathnam(((nmeds>1)+1+(im-1)),1).
            end if.
            compute sobel(im,1) = coeff(2,1).
            compute sobel(im,2)=standerr(2,1).
            compute temp = mnames(1,im).
            compute r2full = 1-(sse/sstm). 
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute summ = {sqrt(r2full), r2full, mse, Fratio, dfnum, dfden, pfr}.
            do if (detail = 1).
              print temp/title = "***************************************************************"+
    "***********"/rlabels = "Outcome:"/format = A8.
              /* do if (cluster = 0) */.
              do if (ws <> 1).
                print summ/title = "Model Summary"/clabels "R", "R-sq", "MSE","F", "df1", "df2", 
    "p"/format = !decimals.    
              end if.             
              /* end if */.
              do if (coeffci = 0).
                compute op = op(:,1:(ncol(op)-2)).
              end if.
              print op/title = "Model"/rnames = xmnm/clabels "coeff", "se", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
              do if (covcoeff=1 and ws <> 1).
                compute hccovtmp=hccov(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(xmnm).
                print hccovtmp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=xmnm/cnames=cnamestp/format=!decimals.
              end if.
              do if (nmods > 0 and nrow(mintkey) > 1).
                print mintkey/title = "Product terms key:"/format = A8.
              end if.
              do if (ws = 1).
                compute wsdf=n-1.
                print wsdf/title=" "/rlabels = "df = "/space=0.
              end if.
            end if.
          end if.
          /* now we create the matrices for indirect effects */.
          compute ymat(1,im) = coeff(2,1).
          do if (ws = 1).
            compute ymat(1,im)=coeff(1,1).
          end if.
          do if (wm = 1).
            compute ymat(2,im) = coeff(4,1).
            do if (zm = 1).
              compute ymat(3,im) = coeff(6,1).
              do if (wzm = 1).
                compute ymat(4, im) = coeff(8,1).
              end if.
            end if.
          end if.
          do if (model = 6).
            compute mmpaths((im+1),1)=coeff((im+1),1).
          end if.
        end loop.
      end if.
      /* estimate model of outcome */.
      loop totlp = 1 to (1+(toteff*(bt = 1))).
        do if (toteff = 1 and totlp =2).
          compute xy = {cons,datayed(:,(3+nmeds):ydatacol)}.
        else.
          compute xy={cons, datayed(:,3:ydatacol)}.
        end if.
        do if (dichy = 1).
          compute pt2 = make(nrow(datayed(:,2)),1,(csum(datayed(:,2))/n)).
          compute LL3 = datayed(:,2)&*ln(pt2)+(1-datayed(:,2))&*ln(1-pt2).
          compute LL3 = -2*csum(LL3).
          compute pt1 = make(n,1,0.5). 
          compute bt1 = make(ncol(xy),1,0). 
          compute LL1 = 0.
          compute xy22=xy.
          loop jjj = 1 to iterate.
            loop ijk=1 to ncol(xy).
              compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
            end loop.
            compute coeff = bt1+inv(t(xy22)*xy)*t(xy)*(datayed(:,2)-pt1).
            compute pt1 = 1/(1+exp(-(xy*coeff))).
            compute itprob = csum((pt1 < .00000000000001) or (pt1 > .99999999999999)).
            do if (itprob = 0).
              compute LL = datayed(:,2)&*ln(pt1)+(1-datayed(:,2))&*ln(1-pt1).
              compute LL2 = -2*csum(ll).
            end if.
            do if (abs(LL1-LL2) < converge).
              break.
            end if.
            compute bt1 = coeff.
            compute LL1 = LL2.
          end loop.
          do if (jjj >= iterate and iterr = 0).
            compute errs = errs+1.
            compute runerrs(errs,1) = 22.
            computer iterr = 1.
          end if.
          loop ijk=1 to ncol(xy).
            compute xy22(:,ijk)=xy(:,ijk)&*pt1&*(1-pt1).
          end loop.
          compute covmat = inv(t(xy22)*xy).
          release xy22.
          do if (totlp <> 2).
            compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
        end if.
        do if (dichy = 0).
          compute invXtX = inv(t(xy)*xy).
          compute coeff =invXtX*t(xy)*datayed(:,2).
          do if (effsize=1 and model > 3 and model < 7 and (ncovs > 0 or clsdmy > 0)).
          compute xysdy = {cons,datayed(:,(4+nmeds):ydatacol)}.
          compute coeffsd =inv(t(xysdy)*xysdy)*t(xysdy)*datayed(:,2).
          compute resid=datayed(:,2)-(xysdy*coeffsd).
          compute sdycov =sqrt(csum(resid&*resid)/(n-ncol(xysdy))).
          compute xvaron=datayed(:,(3+nmeds)).
          compute coeffx2=inv(t(xysdy)*xysdy)*t(xysdy)*xvaron.
          compute residx2=xvaron-(xysdy*coeffx2).
          compute sdxcov =sqrt(csum(residx2&*residx2)/(n-ncol(xysdy))).
          end if. 
          do if (totlp <> 2).
              compute bootcoef={bootcoef,t(coeff(1:(nrow(coeff)-clsdmy),1))}.
          end if.
          do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4 and bt > 1).
            compute resid=datayed(:,2)-xy*coeff.
            compute sse = cssq(resid).
            compute r2full = 1-(sse/sstot).
          end if.
          do if (bt =1).
            compute resid=data(:,2)-xy*coeff.
            compute k3 = nrow(coeff).
            compute sse = cssq(resid).
            compute mse = sse/(n-ncol(xy)).   
            compute hcnot=coeff&/(sqrt(diag(mse*invXtX))).
            do if (hc3 = 1).
              compute h = xy(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xy(i3,:)*invXtX*t(xy(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xy(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xy(:,i3).
              end loop.
            end if.
            do if (hc3 <> 1).
              loop i3=1 to k3.
                compute xy(:,i3) = sqrt(mse)&*xy(:,i3).
              end loop.
            end if.
            compute covmat = (invXtX*t(xy)*xy*invXtX).
          end if.
        end if.
        do if (bt = 1).
          do if (model = 2).
            compute xy2={cons, datayed(:,3:ydatacol)}.
            compute temp = ncol(xy2).
            do if (temp > 6).
              compute xy3=xy2(:,7:temp)}.
            end if.
            compute xy2={xy2(:,1:3), xy2(:,5)}.
            do if (temp > 6).
              compute xy2={xy2, xy3}.
              release xy3.
            end if.
            compute invXtX = inv(t(xy2)*xy2).
            compute coeff2 =invXtX*t(xy2)*datayed(:,2). 
            compute ssem2=cssq(datayed(:,2)-xy2*coeff2).
            release xy2.
          end if.
          compute standerr = sqrt(diag(covmat)).
          do if (totlp =1).
            compute mcmats((nmeds+1):ncol(mcmats),(nmeds+1):ncol(mcmats))=covmat(2:(1+nmeds),2:(1+
    nmeds)).
          end if.
          compute standerr=standerr(1:(nrow(standerr)-clsdmy),1).
          do if (ws=1 and totlp=2).
            compute standerr(1,1)=stddevy/sqrt(n).
          end if.
          compute coeffplt = coeff.
          compute lmat = ident(nrow(coeff)).
          compute lmat = lmat(:,2:ncol(lmat)).
          compute dfnum = nrow(coeff)-1.
          compute dfden = n-dfnum-1.
          compute fratio = (t(t(lmat)*coeff)*inv(t(lmat)*covmat*lmat)*((t(lmat)*coeff)))/dfnum).
          compute coeff=coeff(1:(nrow(coeff)-clsdmy),1).
          do if (totlp=1).
            compute mccoeff((nmeds+1):nrow(mccoeff))=coeff(2:(1+nmeds),1).
          end if.
          compute bbbb=coeff(2,1).
          do if totlp = 1.
            compute deco(1,1)=2+nmeds.
            compute deco = deco(1:decoc,1).
            compute covdirt = make((nrow(covmat)-clsdmy),(ncol(covmat)-clsdmy),0).
            compute covdirt = covmat(deco,:).
            compute covdir = make(nrow(covdirt),ncol(covdirt),0).
            compute covdir = covdirt(:,t(deco)).
            compute deco=coeff(deco,1).
            do if (ttt > 0).
              compute sedir = sqrt(diag(directv*covdir*t(directv))).
              compute directv = directv*deco.
            end if.
            compute sobel(:,3)=coeff(2:(1+nmeds),1).
            compute sobel(:,4) = standerr(2:(1+nmeds),1).
            compute sobel2 = sobel&*sobel.
            do if (varorder <> 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)).
            end if.
            do if (varorder = 2).
              compute sobel(:,2) = sqrt(sobel2(:,1)&*sobel2(:,4)+sobel2(:,3)&*sobel2(:,2)+sobel2(:,
    2)&*sobel2(:,4)).
            end if.
            compute sobel(:,1) = sobel(:,1)&*sobel(:,3).
            compute sobel(:,3) = sobel(:,1)&/sobel(:,2).
            compute sobel(:,4) = 2*(1-cdfnorm(abs(sobel(:,3)))).
          end if.
          do if (dichy = 0).
            compute tratio = coeff&/standerr.
            compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
            do if (ws= 1 and totlp=2).
            compute p = 2*(1-tcdf(abs(tratio), (n-1))).
            compute dfden=n-1.
            end if.
            compute cnms = {"coeff", "se", "t", "p", "LLCI", "ULCI"}.
            compute op = {coeff, standerr, tratio, p}.
          end if.
          do if (dichy = 1).
            compute tratio = (coeff&/standerr).
            compute p = 2*(1-cdfnorm(abs(tratio))).
            compute wald = tratio&*tratio.
            compute cnms = {"coeff", "se", "Z", "p", "LLCI", "ULCI"}.
            compute temp=coeff-abs(xp2)*standerr.
            compute op = {coeff, standerr, tratio, p, temp}.
            compute temp=coeff+abs(xp2)*standerr.
            compute op = {op, temp}.
          end if.
          do if (detail = 1).
            do if (totlp = 2).
              print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
            end if.
            do if (totlp <> 2).
              print yname/title = "**************************************************************"+
    "************"/rlabels = "Outcome:"/format = A8.
            end if.
          end if.
          do if (dichy = 1 and bt = 1 and totlp = 1).
            compute nmsd = {yname, "Analysis"}.
            print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
          end if.
          do if (dichy = 0).
            compute r2full = 1-(sse/ssty).
            compute pfr = 1-fcdf(fratio,dfnum,dfden).
            compute jndf=dfden.
            compute wsdf=dfden.
            compute xd = abs(xp2).
            compute jncrit =  (dfden* (exp((dfden-(5/6))*((xd/(dfden-(2/3)+(.11/dfden)))*
    (xd/(dfden-(2/3)+(.11/dfden)))))-1)).
            compute summ = {sqrt(r2full), r2full, mse, fratio, dfnum, dfden, pfr}. 
            compute temp1=coeff-sqrt(jncrit)*standerr.
            compute temp2=coeff+sqrt(jncrit)*standerr.
            compute op = {coeff, standerr, tratio, p, temp1, temp2}.
            do if (detail = 1).
              /* do if (cluster = 0) */.
              do if (ws <> 1).
              print summ/title = "Model Summary"/clabels = "R", "R-sq", "MSE" "F", "df1", "df2", 
    "p"/format = !decimals.
              /* end if */.
              end if.
            end if.
          end if.
          do if (dichy = 1).
            compute LLdiff = LL3-LL2.
            compute pvalue=1-chicdf(LLdiff,(nrow(coeff)-1)).
            compute mcF = LLdiff/LL3.
            compute cox = 1-exp(-LLdiff/n).
            compute nagel = cox/(1-exp(-(LL3)/n)).
            compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
            do if (detail = 1).
              print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" 
    "p-value" "McFadden" "CoxSnell" "Nagelkrk" "n"/format = !decimals.
            end if.
          end if.
          do if (totlp = 2).
            compute datanmy={"constant"; datanmy((nmeds+2):nrow(datanmy),1)}.
          end if.
          do if (detail = 1).
            do if (coeffci = 0).
              compute op = op(:,1:(ncol(op)-2)).
            end if.
            do if (ws=1).
              compute op2=op(1:(1+nmeds),:).
              compute  datanmy2 = {"c' path"; bpathnam(    (1+(nmeds>1)): 
    ((1+(nmeds>1))+(nmeds-1)),1)}.
              do if (totlp=2).
                compute datanmy2={"c path"}.
                compute op2=op(1,:).
                compute wsdf=n-1.
              end if.
              print op2/title = "Model"/rnames = datanmy2/cnames = cnms/format = !decimals.
              print wsdf/title=" "/rlabels = "df = "/space=0.
            end if.
            do if (ws <> 1).
              print op/title = "Model"/rnames = datanmy/cnames = cnms/format = !decimals.
              do if (covcoeff=1).
                compute covmattp=covmat(1:nrow(op),1:nrow(op)).
                compute cnamestp=t(datanmy).
                print covmattp/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=datanmy/cnames=cnamestp/format=!decimals.
              end if.    
            end if.
          end if.
          do if (ttt = 0 and totlp = 1).
            compute deco = op((nmeds+2),:).
            do if (ws = 1).
              compute deco=op(1,:).
            end if.
          end if.
          do if (ttt = 0 and totlp = 2).
            compute decotot = op(2,:).
            do if (ws = 1).
              compute decotot=op(1,:).
            end if.
          end if.
          do if (nmods > 0 and model > 4 and detail = 1 and nrow(yintkey) > 1)).
            print yintkey/title = "Product terms key:"/format = A8.
          end if.
          do if (nmods > 0 and model < 4 and detail = 1).
            print yintkey2/title = "Product terms key:"/format = A8.
            do if ((model = 1 or model = 2) and dichy = 0).
              compute temp={((op(4,3)**2)*(1-r2full))/dfden, op(4,3)**2, 1, dfden, op(4,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(4,1)**2)*(1-r2full))/dfden.
              end if.
              compute rnms=yintkey2(2,1).
              do if (model = 2).
                 compute temp={temp;((op(6,3)**2)*(1-r2full))/dfden, op(6,3)**2, 1, dfden, op(6,4)}.    
                 do if (hc3=1).
                   compute temp(2,1)=((hcnot(6,1)**2)*(1-r2full))/dfden.
                 end if.
                 compute frat2=(dfden*(r2full-(1-(ssem2/ssty))))/(2*(1-r2full)).
                 do if (hc3=1).
                    compute lmat=make(nrow(coeffplt),2,0).
                    compute lmat(4,1)=1.
                    compute lmat(6,2)=1.
                    compute frat2 = (t(t(lmat)*coeffplt)*inv(t(lmat)*covmat*lmat)*((t(lmat)*
    coeffplt)))/2).
                 end if.
                 compute temp={temp;(r2full-(1-(ssem2/ssty))),frat2,2,dfden,1-fcdf(frat2,2,dfden)}.
                 compute rnms={rnms; yintkey2(3,1);"Both"}.
              end if.
              print temp/title = "R-square increase due to interaction(s):"/rnames=rnms/clabels 
    = "R2-chng", "F", "df1", "df2", "p"/format = !decimals.
            end if.
            do if (model = 3 and dichy = 0).
              compute temp={((op(8,3)**2)*(1-r2full))/dfden, op(8,3)**2, dfden, op(8,4)}.
              do if (hc3=1).
                compute temp(1,1)=((hcnot(8,1)**2)*(1-r2full))/dfden.
              end if.          
              compute rnms=yintkey2(5,1).
              print temp/title = "R-square increase due to three-way "+
    "interaction:"/rnames=rnms/clabels = "R2-chng", "F(1,df2)", "df2", "p"/format = !decimals.
            end if.
          end if.
        end if.
        do if (model = 6 and totlp = 1).
          compute mmpaths(nrow(mmpaths),1)=coeff(nrow(mmpaths),1).
          compute mmpaths(nrow(mmpaths),(2:(nmeds+1)))=t(coeff(2:(nmeds+1),1)).
        end if.
        do if (totlp = 1).
          loop im = 1 to nmeds.
          do if (model < 4).
            compute ymat(1,im) = coeff(3,1)*(1-yes).
            compute ymat(2,im) = coeff(4,1).
            compute cmat(1,im) = covmat(3,3).
            compute cmat(2,im) = covmat(4,4).
            compute cmat(5,im) = covmat(3,4).
            compute jnb1=coeff(3,1).
            compute jnb3=coeff(4,1).
            compute jnsb1=covmat(3,3).
            compute jnsb3=covmat(4,4).
            compute jnsb1b3=covmat(3,4).
            do if (model = 2 or model = 3).
              compute ymat(3,im) = coeff(6,1).
              compute cmat(3,im) = covmat(6,6).
              compute cmat(6,im) = covmat(3,6).
              compute cmat(8,im) = covmat(4,6).
            end if.
            do if (model = 3).
              compute ymat(4,im) = coeff(8,1).
              compute cmat(4,im) = covmat(8,8).
              compute cmat(7,im) = covmat(3,8).
              compute cmat(9,im) = covmat(4,8).
              compute cmat(10,im) = covmat(6,8).
              compute jnb1=coeff(4,1).
              compute jnb3=coeff(8,1).
              compute jnsb1=covmat(4,4).
              compute jnsb3=covmat(8,8).
              compute jnsb1b3=covmat(4,8).
            end if. 
          end if.
          do if (model > 3).
            compute ymat(5,im) = coeff((1+im),1).
          end if.
          do if (xmy = 1).
            compute ymat(6,im) = coeff((2+nmeds+im),1).
          end if.
          do if (vy = 1).
            compute ymat(6,im) = coeff((3+nmeds+im),1).
          end if.
          do if (qy = 1 and vy = 1).
            compute ymat(6,im) = coeff((5+nmeds+((im-1)*2)),1).
            compute ymat(7,im) = coeff((6+nmeds+((im-1)*2)),1).
          end if.
          do if (vqy = 1).
            compute ymat(6,im) = coeff((6+nmeds+((im-1)*3)),1).
            compute ymat(7,im) = coeff((7+nmeds+((im-1)*3)),1).
            compute ymat(8,im) = coeff((8+nmeds+((im-1)*3)),1).
          end if.
          do if (wmy = 1).
            compute ymat(7,im) =  coeff((3+nmeds+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1).
            compute ymat(7,im) =coeff((4+(nmeds*2)+im-wy),1).
          end if.
          do if (wmy = 1 and vy = 1 and wvmy = 1).
            compute ymat(7,im) = coeff((6+(nmeds*2)+((im-1)*2)-wy),1).
            compute ymat(8,im) = coeff((7+(nmeds*2)+((im-1)*2)-wy),1).
          end if.
          do if (wmy = 1 and zmy = 1).
            compute ymat(6,im) = coeff((6-(wzy*3)+(wzm-1)+(nmeds*2)+((im-1)*2)*wzm)+((im-1)*(1-wzm))
    -((zy-wzy)*2),1).
            compute ymat(7,im) = coeff((3-zy+im+nmeds),1).
            do if (wzmy = 1).
              compute ymat(8,im) = coeff((7-(wzy*3)+(nmeds*2)+((im-1)*2)),1).
            end if.
          end if.
          do if (nmods > 0 and model <> 5).
            loop indlp = 1 to nrow(modvals).   
              compute indeff(indlp,1) = csum(ymat(1:4,im)&*vmat(1:4,indlp)).
              do if (model > 6).
                compute indeff(indlp,1) = 
    csum(ymat(1:4,im)&*vmat(1:4,indlp))*csum(ymat(5:8,im)&*vmat(5:8,indlp)).
              end if.
            end loop.
            compute indboot((bt+(im-1)*(boot+1)),:)=t(indeff).
            do if (model = 8 or model = 7).
              compute indbootp(bt,im)=ymat(2,im)*ymat(5,im).
              do if (wdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxw-cminw).
              end if.
            end if.
            do if (model = 14 or model = 15 or model = 74).
              compute indbootp(bt,im)=ymat(1,im)*ymat(6,im).
              do if (wvdich=1).
                compute indbootp(bt,im)=indbootp(bt,im)*(cmaxv-cminv).
              end if.
            end if.
            do if (model = 12).
              compute indbootp(bt,im)=ymat(4,im)*ymat(5,im).
            end if.
            do if ((model = 58 or model = 59) and wvdich=1).
              compute indbootp(bt,im)=(ymat(1,im)*ymat(7,im)*(cmaxw-cminw))+(ymat(2,im)*ymat(5,im)*
    (cmaxw-cminw))+(ymat(2,im)*ymat(7,im)*((cmaxw*cmaxw)-(cminw*cminw))).
            end if.
          end if.
          do if (model = 4 or model = 5).
            compute indboot(bt,im)=csum(ymat(1:4,im)&*vmat(1:4,1))*csum(ymat(5:8,im)&*vmat(5:8,1)).
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute sumind=rsum(indboot(bt,:)).
              do if (im=nmeds).
                compute pmeff(bt,2:(im+1))=indboot(bt,:)/(sumind+coeff((2+nmeds),1)).
              end if.
              compute rmeff(bt,(im+1))=indboot(bt,im)/coeff((2+nmeds),1).
              compute abpseff(bt,(im+1))=indboot(bt,im)/stddevy.
              compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*stddevx.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,(im+1))=indboot(bt,im)/sdycov.
                compute abcseff(bt,(im+1))=abpseff(bt,(im+1))*sdxcov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,1) = r2my-(r2full-r2xy).
                compute abmmr = 1.
                do if (indboot(bt,im) < 0).
                  compute abmmr = cmin(abmm).
                end if.
                do if (indboot(bt,im) > 0).
                  compute abmmr = cmax(abmm).
                end if.
                compute kappa2(bt,1)=indboot(bt,im)/abmmr.
                compute tmp = indboot(bt,im)/abmmr.             
              end if.
            end if.
          end if.
          do if (model = 6). 
            do if (nmeds = 2).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(4,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3).
              compute indboot(bt,3)=mmpaths(3,1)*mmpaths(4,3).
            end if.
            do if (nmeds = 3).
              compute indboot(bt,1) = mmpaths(2,1)*mmpaths(5,2). 
              compute indboot(bt,2) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3).
              compute indboot(bt,3) = mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4).
              compute indboot(bt,4) = mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,5) = mmpaths(3,1)*mmpaths(5,3).
              compute indboot(bt,6) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4).
              compute indboot(bt,7) = mmpaths(4,1)*mmpaths(5,4).
            end if.
            do if (nmeds = 4).
              compute indboot(bt,1)=mmpaths(2,1)*mmpaths(6,2).
              compute indboot(bt,2)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(6,3).
              compute indboot(bt,3)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(6,4).
              compute indboot(bt,4)=mmpaths(2,1)*mmpaths(5,2)*mmpaths(6,5).
              compute indboot(bt,5)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,6)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(5,3)*mmpaths(6,5).
              compute indboot(bt,7)=mmpaths(2,1)*mmpaths(4,2)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,8)=mmpaths(2,1)*mmpaths(3,2)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5)
    .
              compute indboot(bt,9)=mmpaths(3,1)*mmpaths(6,3).
              compute indboot(bt,10)=mmpaths(3,1)*mmpaths(4,3)*mmpaths(6,4).
              compute indboot(bt,11) =mmpaths(3,1)*mmpaths(5,3)*mmpaths(6,5). 
              compute indboot(bt,12) = mmpaths(3,1)*mmpaths(4,3)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,13) = mmpaths(4,1)*mmpaths(6,4).
              compute indboot(bt,14) = mmpaths(4,1)*mmpaths(5,4)*mmpaths(6,5).
              compute indboot(bt,15) = mmpaths(5,1)*mmpaths(6,5).
            end if.
            do if (effsize = 1 and dichy = 0).
              do if (ctot = 0).
                compute ctot=.00000000000001.
              end if.
              compute pmeff(bt,:)=indboot(bt,:)/(rsum(indboot(bt,:))+mmpaths(nrow(mmpaths),1)).
              compute rmeff(bt,:) = indboot(bt,:)/mmpaths(nrow(mmpaths),1).
              compute abpseff(bt,:)=indboot(bt,:)/stddevy.
              compute abcseff(bt,:)=(stddevx*indboot(bt,:))/stddevy.
              do if (ncovs > 0 or clsdmy > 0).
                compute abpseff(bt,:)=indboot(bt,:)/sdycov.
                compute abcseff(bt,:)=(sdxcov*indboot(bt,:))/sdycov.
              end if.
              do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
                compute r245(bt,:) = r2my-(r2full-r2xy).
              end if.
            end if.
          end if.
        end loop.
      end if.
    end loop.  /* end of outcome variable loop */.
    do if (savboot = 1 and bt > 1).
    compute bootstrp((bt-1),:)=bootcoef.
    end if.
  end loop.    /* end of bootstrap loop */.
  do if (savboot = 1).
     compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
     save bootstrp/outfile = *.
  end if.
  release datayed, dat, datamed, data, y, x, datatm, dataty, mints2, cons, mnv, tmp.
  /* Here are MC CI comps */.
  do if (mc > 0).
    compute x1 = sqrt(-2*ln(uniform(mc,nrow(mcmats))))&*cos((2*3.14159265358979)*uniform(mc,
    nrow(mcmats))).
    compute x1=x1*chol(mcmats).
    loop i=1 to nrow(x1).
      compute x1(i,:)=x1(i,:)+t(mccoeff).
    end loop.
    loop i = 1 to nmeds.
      compute indboot(2:nrow(indboot),i)=x1(:,i)&*x1(:,(i+nmeds)).
    end loop.
  end if.
  do if (ttt = 0 and model > 3).
    do if (toteff = 0).
      print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    else.
      print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    end if.
    do if (model <> 74).
      do if (dichy = 0).
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "t", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "t", "p", "LLCI", 
    "ULCI"/format = !decimals.
      else.
        do if (toteff = 1).
          print decotot/title = "Total effect of X on Y"/clabels = "Effect", "SE", "Z", "p", 
    "LLCI", "ULCI"/format = !decimals.
        end if.
        print deco/title = "Direct effect of X on Y"/clabels = "Effect", "SE", "Z", "p", "LLCI", 
    "ULCI" /format = !decimals.
      end if.
    end if.
  end if.
  do if (ttt > 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
    compute clbs = {modvnm2, "Effect", "SE", "t", "p", "LLCI", "ULCI"}.
    compute tratio = (directv&/sedir).
    compute p = 2*(1-tcdf(abs(tratio), (n-ncol(xy)))).
    compute outp = {modvalsd, directv, sedir, tratio, p}.
    do if (dichy = 0).
      compute temp1=directv-sqrt(jncrit)*sedir.
      compute temp2=directv+sqrt(jncrit)*sedir.
      compute outp = {outp, temp1, temp2}.
    end if.
    do if (dichy = 1).
      compute p = 2*(1-cdfnorm(abs(tratio))).
      compute temp = directv-abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute temp = directv+abs(xp2)*sedir.
      compute outp = {outp, temp}.
      compute clbs = {modvnm2, "Effect", "SE", "Z", "p", "LLCI", "ULCI"}.
    end if.
    do if (coeffci = 0).
      compute outp = outp(:,1:(ncol(outp)-2)).
    end if.
    print outp/title = "Conditional direct effect(s) of X on Y at values of the "+
    "moderator(s):"/format = !decimals/cnames = clbs.
  end if.
  do if (nmods > 0 and model <> 5).
    do if (model < 4).
      print/title = "*************************************************************************".
      compute zmat(1,1) = 1*(1-yes).
      compute cfse = make(nrow(modvals),1,0).
      loop #m = 1 to nrow(modvals).
        do if (model = 1).
          compute zmat(2,1)=modvals(#m,1)**2.
          compute zmat(5,1)=2*modvals(#m,1).
        end if.
        do if (model = 2 or model = 3).
          compute zmat(2,1)=modvals(#m,2)**2.
          compute zmat(3,1)=modvals(#m,1)**2.
          compute zmat(4,1)=(modvals(#m,1)**2)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(5,1)=2*modvals(#m,2)*(1-yes).
          compute zmat(6,1)=2*modvals(#m,1)*(1-yes).
          compute zmat(7,1)=2*modvals(#m,1)*modvals(#m,2)*(1-yes).
          compute zmat(8,1)=2*modvals(#m,1)*modvals(#m,2).
          compute zmat(9,1)=2*modvals(#m,1)*(modvals(#m,2)**2)*(1-yes2).
          compute zmat(10,1)=2*(modvals(#m,1)**2)*modvals(#m,2)*(1-yes2).
        end if.
        compute cfse(#m,1)=sqrt(csum(zmat&*cmat)).
      end loop.   
    end if.
    do if (nmods > 0).
      compute clbs = {modvnm, "Effect"}.
      loop im = 1 to nmeds.
        compute obs = t(indboot(1+(im-1)*(boot+1),:)).
        compute outp = {modvals, obs}.
        do if (model < 4).
          compute tstat = obs&/cfse.
          do if (dichy = 0).
            compute pval = 2*(1-tcdf(abs(tstat), (n-ncol(xy)))).
            compute temp=obs-sqrt(jncrit)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp=obs+sqrt(jncrit)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "t", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
	         end if.
          do if (dichy = 1).
            compute pval = 2*(1-cdfnorm(abs(tstat))).
            compute temp = obs-abs(xp2)*cfse.
            compute outp = {outp, cfse, tstat, pval, temp}.
            compute temp = obs+abs(xp2)*cfse.
            compute outp = {outp, temp}.
            compute clbs = {clbs, "se", "Z", "p", "LLCI", "ULCI"}.
            compute jnclbs=clbs.
          end if.
        end if.
        do if (boot > 0).
          compute ones = make(boot,1,1).
          compute estmte=indboot((1+(im-1)*(boot+1)),:).
          compute indboot2 = indboot((2+(im-1)*(boot+1)):(1+(im-1)*(boot+1)+boot),:).
          compute mnind = t(csum(indboot2)/boot).
          compute stdind=t(sqrt((cssq(indboot2)-((csum(indboot2)&**2)/boot))/(boot-1))).
          compute llci=make(1,ncol(indboot2),-999).
          compute ulci=make(1,ncol(indboot2),-999).
          loop #e = 1 to ncol(indboot2).
             bcboot databcbt = indboot2(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
             compute llci(1,#e)=llcit.
             compute ulci(1,#e)=ulcit.
             do if (badlo = 1 and llcit <> priorlo).
               compute badend={badend, llcit}.
               compute priorlo = llcit.
             end if.
             do if (badhi = 1 and ulcit <> priorhi).
               compute badend={badend, ulcit}.
               compute priorhi = ulcit.
             end if.
          end loop.
          compute outp = {modvals, obs, stdind, t(llci), t(ulci)}.
          compute clbs = {modvnm, "Effect", "Boot SE", "BootLLCI", "BootULCI"}.
        end if.
        compute mtemp = mnames(1, im).
        compute rlbs = make(nrow(modvals),1,mnames(1,im)).
        do if (model < 4).
          do if (coeffci = 0).
            compute outp = outp(:,1:(ncol(outp)-2)).
          end if.
          do if (yes=0).
            print outp/title = "Conditional effect of X on Y at values of the moderator(s):"/cnames 
    = clbs/format = !decimals.
          else.
            compute outp=outp(:,3:ncol(outp)).
            compute clbs=clbs(:,3:ncol(clbs)).
            compute clbs(1,1)="Contrast".
            print outp/title = "Contrast of conditional effects of X on Y"/cnames = clbs/format = 
    !decimals.
          end if.
        end if.
        do if (model > 5 and mod74dic <> 1).
          do if (im = 1).
            print/title = "Conditional indirect effect(s) of X on Y at values of the moderator(s):".    
          end if.
          print outp/title = "Mediator"/rnames = rlbs/cnames = clbs/format = !decimals.
        else if (mod74dic = 1 and model = 74).
          do if (im = 1).
            print/title = "Indirect effect(s) of X on Y:".
            compute clbs3=clbs(1,2:ncol(clbs)).
          end if.
          compute outp3=outp(1,2:ncol(outp)).
          print outp3/title = "Mediator"/rnames = rlbs/cnames = clbs3/format = !decimals.
        end if.
      end loop.
      loop i = notes to 1 by -1.
      do if (note(i,1) = 4 and yes=0).
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 5 and yes=0).
        print/title = "Values for quantitative moderators are the mean and plus/minus one SD "+
    "from mean.".
        print/title = "Values for dichotomous moderators are the two values of the "+
    "moderator."/space=0.
      end if.
      do if (note(i,1) = 14 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      below the mean was replaced with the minimum because one SD below "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      do if (note(i,1) = 15 and yes=0).
        print/title = "NOTE: For at least one moderator in the conditional effects table above, "+
    "one SD ".
        print/title = "      above the mean was replaced with the maximum because one SD above "+
    "the mean"/space=0.
        print/title = "      is outside of the range of the data."/space=0.   
      end if.
      end loop.
      do if (model = 3 and yes=0).
        compute jnvals=make(nrow(matw),7,0).
        compute jnvals(:,1)=matw.
                compute jnvals(:,2)=jnb1+jnb3*jnvals(:,1).
                compute jnvals(:,3)=sqrt(jnsb1+2*jnvals(:,1)*jnsb1b3+(jnvals(:,1)&*jnvals(:,1))*
    jnsb3).
                compute jnvals(:,4)=jnvals(:,2)&/jnvals(:,3).
                do if (dichy = 0).
                  compute jnvals(:,5)=2*(1-tcdf(abs(jnvals(:,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(:,5)=2*(1-cdfnorm(abs(jnvals(:,4)))).
                end if.
                compute jnvals(:,6)=jnvals(:,2)-sqrt(jncrit)&*jnvals(:,3).
                compute jnvals(:,7)=jnvals(:,2)+sqrt(jncrit)&*jnvals(:,3).
               compute clbs={clbs(:,1),clbs(:,3:ncol(clbs))}.
               do if (coeffci = 0).
                 compute jnvals =jnvals(:,1:(ncol(jnvals)-2)).
              end if.
              print jnvals/title = "Conditional effect of X*M interaction at values of W:"/cnames 
    =clbs/format = !decimals.
      end if.
      do if (jn = 1 and (model = 1 or model = 3) and jndich = 0 and yes=0).
        print/title = "********************* JOHNSON-NEYMAN TECHNIQUE **************************".
        compute ajn =(jncrit*jnsb3)-(jnb3*jnb3).
        compute bjn = 2*((jncrit*jnsb1b3)-(jnb1*jnb3)).
        compute cjn = (jncrit*jnsb1)-(jnb1*jnb1).
        compute radarg = (bjn*bjn)-(4*ajn*cjn).
        compute den = 2*ajn.
        compute nrts = 0.
        do if (radarg >= 0 and den <> 0).
          compute x21 = (-bjn+sqrt(radarg))/den.
          compute x22 = (-bjn-sqrt(radarg))/den.
          compute roots = 0.
          do if (x21 >= jnmin and x21 <= jnmax).
            compute nrts = 1.
            compute roots = {roots; x21}.
          end if.
          do if (x22 >= jnmin and x22 <= jnmax).
            compute nrts = nrts + 1.
            compute roots = {roots; x22}.
          end if.
          compute roots={roots,make(nrow(roots),2,0)}.
          compute modtemp=m.
          do if (model=3).
            compute modtemp=w.
          end if.
          do if (nrts > 0).                           
            compute roots = roots(2:nrow(roots),1:3).
            compute roots(1,2)=(csum(modtemp < roots(1,1))/n)*100.
            compute roots(1,3)=(csum(modtemp > roots(1,1))/n)*100.
            do if (nrow(roots)=2).
              compute roots(2,2)=(csum(modtemp < roots(2,1))/n)*100.
              compute roots(2,3)=(csum(modtemp > roots(2,1))/n)*100.
            end if.
            print roots/title = "Moderator value(s) defining Johnson-Neyman significance "+
    "region(s)"/clabels = "Value", "% below", "% above"/format = !decimals.
            compute jnvals=make((21+nrts),7,0).
            loop i= 0 to 20.
              compute jnvals((i+1),1)=jnmin+(i*((jnmax-jnmin)/20)).
            end loop.
            loop i = 1 to nrts.
              loop j = 2 to nrow(jnvals).
                do if ((roots(i,1) > jnvals((j-1),1)) and (roots(i,1) < jnvals(j,1))).
                  compute jnvals((j+1):(21+i),1)=jnvals(j:(20+i),1).
                  compute jnvals(j,1)=roots(i,1).
                end if.
              end loop.
            end loop. 
            loop i = 1 to nrow(jnvals).
                compute jnvals(i,2)=jnb1+jnb3*jnvals(i,1).
                compute jnvals(i,3)=sqrt(jnsb1+2*jnvals(i,1)*jnsb1b3+(jnvals(i,1)*jnvals(i,1))*
    jnsb3).
                compute jnvals(i,4)=jnvals(i,2)/jnvals(i,3).
                do if (dichy = 0).
                  compute jnvals(i,5)=2*(1-tcdf(abs(jnvals(i,4)), jndf)).
                end if.
                do if (dichy = 1).
                  compute jnvals(i,5)=2*(1-cdfnorm(abs(jnvals(i,4)))).
                end if.
                compute jnvals(i,6)=jnvals(i,2)-sqrt(jncrit)*jnvals(i,3).
                compute jnvals(i,7)=jnvals(i,2)+sqrt(jncrit)*jnvals(i,3).
            end loop. 
            do if (model = 1).
              print jnvals/title = "Conditional effect of X on Y at values of the moderator "+
    "(M)"/cnames =jnclbs/format = !decimals.
            end if.
            do if (model = 3).
              compute jnclbs={jnclbs(:,1),jnclbs(:,3:ncol(jnclbs))}.
              print jnvals/title = "Conditional effect of X*M on Y at values of the moderator "+
    "(W)"/cnames =jnclbs/format = !decimals.
            end if.
          end if.
          do if (nrts = 0).
            print/title = "There are no statistical significance transition points within the "+
    "observed".
            print/title = "range of the moderator."/space=0.
          end if.
        else.
           print/title = "There are no statistical significance transition points within the "+
    "observed".
           print/title = "range of the moderator."/space=0.
        end if.
      end if.
    end if.
    /* print data for plotting */.
    do if (model < 4 and (plot = 1)).
      compute dataplot = make((nrow(modvals)*nrow(matx)),(ncol(modvals)+1),0).
      compute tmp = 1.
      loop i = 1 to nrow(modvals).
        loop j = 1 to nrow(matx).
          compute dataplot(tmp,:)={matx(j,1), modvals(i,:)}.
          compute tmp=tmp+1.
        end loop.
      end loop.
      compute dataplot = {dataplot, make(nrow(dataplot),(1+dichy),0)}.
      compute dataplo2 = make(nrow(dataplot),1,1).
      do if (model = 1).
        compute dataplo2 = {dataplo2, dataplot(:,2), dataplot(:,1), (dataplot(:,1)&*dataplot(:,2))}.    
      end if.
      do if (model = 2 or model = 3).
        compute dataplo2 = {dataplo2, dataplot(:,3), dataplot(:,1),  
    (dataplot(:,1)&*dataplot(:,3)), dataplot(:,2), (dataplot(:,1)&*dataplot(:,2))}.
        do if (model = 3).
          compute dataplo2 = {dataplo2, (dataplot(:,2)&*dataplot(:,3)), 
    (dataplot(:,1)&*dataplot(:,2)&*dataplot(:,3))}.
        end if.
      end if.
      loop i = 1 to nrow(dataplot).
        compute tmp=dataplo2(i,:).
        do if (ncovs > 0).
          compute tmp = {tmp, covmeans}.
        end if.
        do if (cluster > 0).
          compute tmp = {tmp, cldmeans}.
        end if.
        compute dataplot(i,(ncol(dataplot)-(dichy)))=tmp*coeffplt.
        do if (dichy = 1).
          compute dataplot(i,(ncol(dataplot)))=exp(tmp*coeffplt)/(1+exp(tmp*coeffplt)).
        end if.
      end loop.
      compute clbs = {xname, modvnm, "yhat"}.
      do if (dichy = 1).
        compute clbs = {xname, modvnm, "ln(odds)", "prob"}.
      end if.
      print/title = "**************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !w, " ", !m, " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line30 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, "/PANEL ROWVAR=", !w, ".").
      !let !line36 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line42 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line43 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line37 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      !let !line38 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, "/PANEL ROWVAR=", 
    !w, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (model = 1).
        do if (dichy = 0).
          print/title = !quote(!line0)/space=1.
          else if (dichy = 1).
          print/title = !quote(!line02)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich = 0).
            print/title = !quote(!line3)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line30)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich = 0).
            print/title = !quote(!line32)/space=0.
            print/title = !quote(!line33)/space=0.
          end if.
          do if (xdich <> 0).
            print/title = !quote(!line42)/space=0.
            print/title = !quote(!line43)/space=0.
          end if.
        end if.
      end if.
      do if (model = 2 or model = 3).
        do if (dichy = 0).
          print/title = !quote(!line01)/space=1.
        end if.
        do if (dichy = 1).
          print/title = !quote(!line03)/space=1.
        end if.
        print/title = !quote(!line1)/space=0.
        print dataplot/title=" "/format = !decimals/space=0.
        print/title = !quote(!line2)/space=1.
        do if (dichy = 0).
          do if (xdich=0).
            print/title = !quote(!line31)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line36)/space=0.
          end if.
        end if.
        do if (dichy = 1).
          do if (xdich=0).
            print/title = !quote(!line34)/space=0.
            print/title = !quote(!line35)/space=0.
          end if.
          do if (xdich<>0).
            print/title = !quote(!line37)/space=0.
            print/title = !quote(!line38)/space=0.
          end if.
        end if.
      end if.
      do if (ncovs > 0).
        print/title = "* Estimates are based on setting covariates to their sample means.".
      end if.
    end if.
   end if.
  do if (model = 7 or model = 8 or model = 12 or model = 14 or model = 15 or (model = 74 and 
    mod74dic=0) or ((model = 58 or model = 59) and wvdich=1)).
    compute obsprod = t(indbootp(1,:)).
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute estmte=indbootp(1,:).
      compute indbootp = indbootp(2:(boot+1),:).
      compute mnindp = t(csum(indbootp)/boot).
      compute stdindp=t(sqrt((cssq(indbootp)-((csum(indbootp)&**2)/boot))/(boot-1))).
      compute llcip=make(1,ncol(indbootp),-999).
      compute ulcip=make(1,ncol(indbootp),-999).
      loop #e = 1 to ncol(indbootp).
        bcboot databcbt = indbootp(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llcip(1,#e)=llcit.
        compute ulcip(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute outp = {obsprod, stdindp, t(llcip), t(ulcip)}.
      compute clbs = {"Effect", "SE(Boot)", "BootLLCI", "BootULCI"}.
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print outp/title = "Mediator"/cnames = clbs/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        compute clbs3 = {"Index", "SE(Boot)", "BootLLCI", "BootULCI"}.
        print outp/title = "Mediator"/cnames = clbs3/rnames = mnames/format = !decimals.
          do if (wvdich=1).
            print/title = "When the moderator is dichotomous, this is a test of equality of the".
            print/title = "conditional indirect effects in the two groups."/space=0.
          end if.
      end if.
    end if.
    do if (boot = 0).
      do if (model = 8 or model = 12).
        print/title="-----".
        print/title = "Indirect effect of highest order product:"/space=0.
        print obsprod/title = "Mediator"/clabels = "Effect"/rnames = mnames/format = !decimals.
      end if.
      do if (model <> 12).
        print/title = "******************** INDEX OF MODERATED MEDIATION ************************".
        print obsprod/title = "Mediator"/clabels = "Index"/rnames = mnames/format = !decimals.
      end if.
    end if.
  end if.
  compute conmake=0.
  compute concols = 0.
  do if ((model > 3 and model < 7) and (contrast > 0) and nmods = 0 and nmeds > 1 ).
    compute concols = ((ncol(indboot)*(ncol(indboot)-1))/2).
    compute indcon=make(nrow(indboot),concols,-999).
    compute conkey = {"  ", "  ", "  "}.
    compute temp=1.
    compute conmake=1.
    loop i = 1 to (ncol(indboot)-1).
      loop j = (i+1) to (ncol(indboot)).
        compute indcon(:,temp)=indboot(:,i)-indboot(:,j).
        do if (contrast = 2).
          compute indcon(:,temp)=abs(indboot(:,i))-abs(indboot(:,j)).
        end if.
        do if (model <> 6).
          compute conkey={conkey; mnames(1,i), "minus", mnames(1,j)}.
        end if.
        do if (model = 6).
          compute conkey={conkey; indlbl2(i,1), "minus", indlbl2(j,1)}.
        end if.
        compute temp=temp+1.
      end loop.
    end loop.
  end if.
  do if (model = 4 or model = 5).
    compute clbs = {"Effect"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obs = {csum(obs); obs}.
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute rlbs = {rlbs; cntname(1:ncol(indcon),1)}.
    end if.
    compute outp = obs.
    compute outp2=outp.
    do if (effsize = 1 and dichy = 0).
      compute pmeff(:,1)=rsum(pmeff(:,2:ncol(pmeff))).
      compute rmeff(:,1)=rsum(rmeff(:,2:ncol(rmeff))).
      compute abpseff(:,1)=rsum(abpseff(:,2:ncol(abpseff))).
      compute abcseff(:,1)=rsum(abcseff(:,2:ncol(abcseff))).
      compute eff = {pmeff, rmeff, abpseff, abcseff}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute eff = {eff, r245, kappa2}.
        compute r245obs = {r245(1,1);r245(1,1)}.
        compute kappa2ob = kappa2(1,1).
      end if.
      compute pmobs = t(pmeff(1,1:(nmeds+1))).
      compute rmobs = t(rmeff(1,1:(nmeds+1))).
      compute psobs = t(abpseff(1,1:(nmeds+1))).
      compute csobs = t(abcseff(1,1:(nmeds+1))).
      do if (contrast = 0).
         compute outp2 = {obs, psobs, csobs, pmobs, rmobs}.
      end if.
      do if (contrast > 0).
         compute obs2=obs(1:nrow(psobs),:).
         compute outp2 = {obs2, psobs, csobs, pmobs, rmobs}.
      end if.
      compute clbs = {"ab", "ab_ps", "ab_cs", "ab/c", "ab/c'"}.
      do if (nmeds = 1 and ncovs = 0 and cluster = 0 and model = 4).
        compute outp2 = {outp2, r245obs, (obs/abmmr)}.
        compute clbs = {clbs, "R-sq_med", "kappa2"}.
      end if.
    end if.
    do if (boot = 0 and mc = 0).
      do if (nmeds = 1).
        compute outp2 = outp2(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      do if (model=5).
        compute outp2=outp2(:,1:3).
      end if.
      print outp2/title = "Indirect effect(s) of X on Y"/rnames = rlbs/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1 and nmeds > 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute indboot = {rsum(indboot), indboot}.
      compute bootsz=boot.
      do if (mc > 0).
        compute bootsz=mc.       
      end if.
      compute ones = make(bootsz,1,1).
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(bootsz+1),:).
      compute mnind = t(csum(indboot)/bootsz).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/bootsz))/(bootsz-1))).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt=indboot(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      do if (effsize = 1 and dichy = 0 and boot > 0).
        compute estmte=eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt=eff(:,#e)/estmte = (estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
      end if.
    end if.
    do if (boot > 0 or mc > 0).
      compute outp = {obs, stdind, t(llci), t(ulci)}.
      do if (nmeds = 1).
        compute outp = outp(2,:).
        compute rlbs = rlbs(2,1).
      end if.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      do if (mc > 0).
        compute clbs = {"Effect", "MC SE", "MC LLCI", "MC ULCI"}.
      end if.
      print outp/title = "Indirect effect of X on Y"/rnames = rlbs/cnames = clbs/format = !decimals.    
      do if (dichy = 0 and effsize = 1 and boot > 0).
        compute outp = {psobs, stdindf((2*(nmeds+1)+1) :(3*(nmeds+1)),1), 
    t(llcif(1,(2*(nmeds+1)+1):(3*(nmeds+1)))), t(ulcif(1,(2*(nmeds+1)+1):(3*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.
        print outp/title = "Partially standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.
        compute outp = {csobs, stdindf((3*(nmeds+1)+1): (4*(nmeds+1)),1),  
    t(llcif(1,(3*(nmeds+1)+1):(4*(nmeds+1)))), t(ulcif(1,(3*(nmeds+1)+1):(4*(nmeds+1))))}.
        do if (nmeds = 1).      
          compute outp = outp(2,:).
        end if.      
        print outp/title = "Completely standardized indirect effect of X on Y"/rnames = rlbs/cnames 
    = clbs/format = !decimals.    
        compute outp = {pmobs, stdindf(1:(nmeds+1),1),t(llcif(1,1:(nmeds+1))), 
    t(ulcif(1,1:(nmeds+1))) }.
        do if (model = 4).
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          compute outp = {rmobs, stdindf(((nmeds+1)+1) :(2*(nmeds+1)) ,1), 
    t(llcif(1,((nmeds+1)+1):(2*(nmeds+1)))), t(ulcif(1,((nmeds+1)+1):(2*(nmeds+1))))}.
          do if (nmeds = 1).      
            compute outp = outp(2,:).
          end if.      
          print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          do if (nmeds = 1 and cluster = 0 and ncovs = 0 and model = 4).      
            compute r245obs = r245obs(1,1).
            compute outp = {r245obs, stdindf((4*(nmeds+1)+1) :(4*(nmeds+1)+1),1), 
    t(llcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1))), t(ulcif(1,(4*(nmeds+1)+1):(4*(nmeds+1)+1)))}.    
            print outp/title = "R-squared mediation effect size (R-sq_med)"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
            compute outp = {kappa2ob, stdindf(nrow(stdindf),1),llcif(1,ncol(llcif)), 
    ulcif(1,ncol(ulcif))}.    
            print outp/title = "Preacher and Kelley (2011) Kappa-squared"/rnames = rlbs/cnames = 
    clbs/format = !decimals.
          end if.  
        end if.
      end if.
    end if.
    do if (normal = 1).
      do if (nmeds = 1).
        print sobel/title = "Normal theory tests for indirect effect"/clabels = "Effect", "se", 
    "Z", "p"/format = !decimals.
      end if.
      do if (nmeds > 1).
        compute rlbs2 = rlbs(2:nrow(rlbs),1).
        print sobel/title = "Normal theory tests for specific indirect effects"/rnames = 
    rlbs2/clabels = "Effect", "se", "Z", "p"/format = !decimals.
      end if.
    end if.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  do if (model = 6).
    compute clbs = {"eff"}.
    compute rlbs = {"TOTAL"; t(mnames)}.
    compute obs = t(indboot(1,:)).
    compute obsfsum=csum(obs).
    compute obs = {obsfsum; obs}.
    compute indlbl = indlbl(1:nrow(obs),1).
    do if (conmake = 1).
      compute obs={obs; t(indcon(1,:))}.
      compute indlbl = {indlbl; cntname(1:ncol(indcon),1)}.
    end if.
    compute obs2=obs.
    do if (boot = 0).
      do if (dichy = 0 and effsize = 1).
        compute clbs = {"eff", "eff_ps", "eff_cs", "eff/c", "eff/c'"}.
        do if (ncovs = 0).
          compute obs = {obs, obs/stddevy, (obs*stddevx/stddevy), obs/ctot, 
    obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        do if (ncovs > 0 or clsdmy > 0).
          compute obs = {obs, obs/sdycov, (obs*sdxcov/sdycov), 
    obs/(obsfsum+mmpaths(nrow(mmpaths),1)), obs/mmpaths(nrow(mmpaths),1)}.
        end if.
        compute obs2=obs.
        do if (contrast > 0).
          compute obs2 = obs(1:(nrow(obs)-concols),:).
        end if.
      end if.
      print obs2/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (contrast > 0 and effsize = 1).
        compute outp2=t(indcon(1,:)).
        compute rlbs2 = cntname(1:ncol(indcon),1).
        print outp2/title = "Contrast(s) between indirect effects"/rnames = 
    rlbs2/cnames=clbs/format = !decimals.
      end if.
    end if.
    do if (boot > 0).
      compute ones = make(boot,1,1).
      compute indboot = {rsum(indboot), indboot}.
      do if (conmake = 1).
        compute indboot={indboot, indcon}.
      end if.
      compute estmte=indboot(1,:).
      compute indboot = indboot(2:(boot+1),:).
      compute mnind = t(csum(indboot)/boot).
      compute stdind=t(sqrt((cssq(indboot)-((csum(indboot)&**2)/boot))/(boot-1))).
      compute temp = nrow(indboot).
      compute llci=make(1,ncol(indboot),-999).
      compute ulci=make(1,ncol(indboot),-999).
      loop #e = 1 to ncol(indboot).
        bcboot databcbt = indboot(:,#e)/estmte=(estmte(1,#e)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,#e)=llcit.
        compute ulci(1,#e)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute obs = {obs, stdind, t(llci), t(ulci)}.
      compute clbs = {"Effect", "Boot SE", "BootLLCI", "BootULCI"}.
      print obs/title = "Indirect effect(s) of X on Y"/rnames = indlbl/cnames = clbs/format = 
    !decimals.
      do if (effsize = 1 and dichy = 0).
        compute indboot=indboot(:,1:(ncol(indboot)-concols)).
        compute eff = {rsum(abpseff), abpseff, rsum(abcseff), abcseff, rsum(pmeff), pmeff, 
    rsum(rmeff), rmeff}. 
        compute effobs = eff(1,:).
        compute eff = eff(2:nrow(eff),:).
        compute stdindf=t(sqrt((cssq(eff)-((csum(eff)&**2)/boot))/(boot-1))).
        compute llcif=make(1,ncol(eff),-999).
        compute ulcif=make(1,ncol(eff),-999).
        loop #e = 1 to ncol(eff).
          bcboot databcbt =eff(:,#e)/estmte = (effobs(1,#e)*bconoff)+(9999*(1-bconoff)).
          compute llcif(1,#e)=llcit.
          compute ulcif(1,#e)=ulcit.
          do if (badlo = 1 and llcit <> priorlo).
            compute badend={badend, llcit}.
            compute priorlo = llcit.
          end if.
          do if (badhi = 1 and ulcit <> priorhi).
            compute badend={badend, ulcit}.
            compute priorhi = ulcit.
          end if.
        end loop.
        compute temp2 = stdindf(1:ncol(indboot),1).
        compute temp3 = effobs(:,1:ncol(indboot)).
        compute templow = llcif(1,1:ncol(indboot)).
        compute temphi = ulcif(1,1:ncol(indboot)).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Partially standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((ncol(indboot)+1):(2*ncol(indboot)),1).
        compute temp3 = effobs(:,(ncol(indboot)+1):(2*ncol(indboot))).
        compute templow = llcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute temphi = ulcif(1,(ncol(indboot)+1):(2*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Completely standardized indirect effect of X on Y"/cnames = clbs/rnames 
    = indlbl/format = !decimals.
        compute temp2 = stdindf((2*(ncol(indboot))+1):(3*ncol(indboot)),1).
        compute temp3 = effobs(:,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute templow = llcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute temphi =  ulcif(1,(2*(ncol(indboot))+1):(3*ncol(indboot))).
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to total effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
        compute temp = eff(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temp2 = stdindf((3*(ncol(indboot))+1):(4*ncol(indboot)),1).
        compute temp3 = effobs(:,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute templow = llcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))).
        compute temphi = ulcif(1,(3*(ncol(indboot))+1):(4*ncol(indboot))). 
        compute outp = {t(temp3), temp2,t(templow), t(temphi)}.
        print outp/title = "Ratio of indirect to direct effect of X on Y"/rnames = indlbl/cnames = 
    clbs/format = !decimals.
      end if.
    end if.
    do if (nmeds = 2).
      compute effkey = {xname, "->", mnames(1,1), "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->", mnames(1,1), "->", mnames(1,2), "->", yname}.
      compute effkey = {effkey; xname, "->", mnames(1,2), "->", yname, " ", " "}.
      compute effkey = {indlbl(2:4,1), effkey}.
    end if.
    do if (nmeds = 3).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " "}.
      compute effkey = {indlbl(2:8,1), effkey}.
    end if.
    do if (nmeds = 4).
      compute effkey = {xname, "->" , mnames(1,1), "->", yname, " ", " ", " ", " ", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,1), "->", mnames(1,2), "->", mnames(1,3), 
    "->", mnames(1,4), "->", yname}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,2), "->", mnames(1,3), "->", mnames(1,4), 
    "->", yname, " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {effkey; xname, "->" , mnames(1,3), "->", mnames(1,4), "->", yname, " ", " "+
    "", " ", " "}.
      compute effkey = {effkey; xname, "->" , mnames(1,4), "->", yname, " ", " ", " ", " ", " ", " "+
    ""}.
      compute effkey = {indlbl(2:16,1), effkey}.
    end if.
    print effkey/title = "Indirect effect key"/format = A8.
    do if (conmake = 1).
      compute conkey=conkey(2:nrow(conkey),:).
      compute conlbs = cntname(1:ncol(indcon),1).
      print conkey/title = "Specific indirect effect contrast definitions"/rnames = conlbs/format = 
    A10.
    end if.
  end if.
  end if.
 DO IF (medte2=1).
    compute name=t(mdlnms2(2:nrow(mdlnms2),1)).
compute tmpdat=datayed(:,2:ncol(datayed)).
compute dd=tmpdat(:,1).
compute dd={dd,tmpdat(:,(2+nmeds)),tmpdat(:,2:(1+nmeds))}.
compute stratify=0.
do if (ncovs > 0).
    compute cname = cnames.
    compute dd={dd,tmpdat(:,(3+nmeds):(2+nmeds+ncovs))}.
end if.
compute method = 1.
compute samples = boot.
do if (mc > 0).
  compute samples = mc.
  compute method=2.
end if.
compute omnibus = 1.
compute booterr = 0.
do if (mcx < 0 or mcx > 4).
   compute mcx = 1.
end if.
compute ovals = ncol(design(dd(:,1))).
compute nte = notes.
compute criterr = 0.
compute xskip = 0.
compute dumok = 0.
compute badend = 0.
compute alpha2 = (1-(conf/100))/2.
compute y5=sqrt(-2*ln(alpha2)).
compute xp2=-(y5+((((y5*p4+p3)*y5+p2)*y5+p1)*y5+p0)/((((y5*q4+q3)*y5+q2)*y5+q1)*y5+q0)).
compute priorlo = -9999999.
compute priorhi = 9999999.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute reps = samples.
do if (contrast = 1).
  compute note(nte,1) = 26.
  compute nte=nte+1.
end if.
do if (effsize = 1).
  compute note(nte,1) = 27.
  compute nte=nte+1.
end if.
do if (normal = 1).
  compute note(nte,1) = 28.
  compute nte=nte+1.
end if.
do if (method = 2).
  do if (omnibus = 1).
    compute note(nte,1) = 25.
    compute nte=nte+1.
  end if.
  compute samples = 0.
  do if (reps < 1000).
     compute reps = 1000.
  end if.
end if.
do if (method = 1).
   compute samples  = reps.
end if.
compute temp = ncol(dd).
compute n = nrow(dd).
compute ones = make(n,1,1).
compute temp=dd(:,2).
compute nxd = ncol(temp).
compute nx = nxd.
do if (nx = 1 and mcx > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,2)),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,2)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute mnvls = cmin(t(nnvls)).
  do if (mnvls < 2).
    compute errs = errs + 1.
    compute runerrs(errs,1) = 37.
    compute criterr = 1.
  end if.
  do if (nvls > 9).
    compute errs = errs+1.
    compute runerrs(errs,1) = 36.
    compute criterr = 1.
  end if.
  do if (criterr = 0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,2).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,2) <> nnvls((temp-1),1)).
         compute nnvls(temp,1)=dd(i,2).
         compute temp = temp+1.
      end if.
    end loop.
    do if (mcx > 0).
      compute x = dummy(:,2:ncol(dummy)).
      compute nx = ncol(x).
      compute minus1 = make(1,ncol(x),-1).
      compute xdes=make((nx+1),3,0).
      compute xdes(1,1)=dd(1,2).
      compute xdes(1,2)=1.
      compute temp = 2.
      loop k = 2 to n.
        do if (dd(k,2) <> dd((k-1),2)).
          compute xdes(temp,2) = k.
          compute xdes(temp,1) = dd(k,2).
          compute xdes((temp-1),3) = k-1.
          compute temp=temp+1.
        end if.
      end loop.
      compute xdes((temp-1),3)=n.
      compute xdes = {xdes, (xdes(:,3)-xdes(:,2)+1)}.
      do if (stratify <> 0).
        compute note(nte,1) = 7.
        compute nte=nte+1.
      end if.
      do if (mcx = 4).
        loop k = 1 to n.
          do if (rsum(x(k,:)) = 0).
            compute x(k,:) = minus1.
          end if.
        end loop.
      end if.
      do if (mcx = 2 or mcx = 3).
         compute note(nte,1)=4.
         compute nte = nte+1.
         loop k = 1 to n.
           do if (rsum(x(k,:)) > 0).
               loop i = 1 to ncol(x).
                   do if (x(k,i) = 0).
                       compute x(k,i) = 1.
                   else.
                        break.
                   end if.
                end loop.
           end if.
         end loop.
         do if (mcx = 3).
           compute conmat1={-8,1,1,1,1,1,1,1,1;
                                              0,-7,1,1,1,1,1,1,1;
                                              0,0,-6,1,1,1,1,1,1;
                                              0,0,0,-5,1,1,1,1,1;
                                              0,0,0,0,-4,1,1,1,1;
                                              0,0,0,0,0,-3,1,1,1;
                                              0,0,0,0,0,0,-2,1,1;
                                              0,0,0,0,0,0,0,-1,1}.
           loop i = 1 to 8.
             compute conmat1(i,:)=conmat1(i,:)/(10-i).
           end loop.
           compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
           loop k=1 to n.
              compute x(k,:)=conmat1((rsum(x(k,:))+1),:). 
           end loop.
        end if.
      end if.
      compute xdata = x.
      compute xskip = 1.
      compute xname = ddd(1,1:nx).
      compute indlbs = t(xname).
      compute dummat = make((nx+1),nx,0).
      compute dummat((2:nrow(dummat)),:)=ident(nx).
      do if (mcx = 4).
         compute dummat(1,:) = minus1.
      end if.
      do if (mcx = 2).
         loop i = 2 to nrow(dummat).
            loop j = 1 to (i-1).
               compute dummat(i,j) = 1.
            end loop.
          end loop.
      end if.
      do if (mcx = 3).
        compute dummat=conmat1.
      end if.
      compute dummat={nnvls, dummat}.
      compute xname2={name(1,2), xname}.   
    end if.
  end if.
end if.
compute nm=nmeds.
compute r2b = make(nm,1,-999).
compute indeff = make((nx*nm),1,-999).
compute cov = make((nx*nm+nm),(nx*nm+nm),0).
compute mns=make(nrow(cov),1,0).
compute ny = ncol(dd)-nm-nxd-ncovs.
compute y = dd(:,1).
compute ydata = y.
compute yname = name(1,1).
do if (xskip = 0).
  compute x = dd(:,2:(1+nxd)).
  compute xdata = x.
  compute xname = name(1,2:(1+nxd)).
  compute indlbs = t(xname).
end if.
compute m = dd(:,(2+nxd):(1+nxd+nm)).
compute mdata = m.
compute mlab = {"   M1 = "; "   M2 ="; "   M3 ="; "   M4 ="; "   M5 ="; "   M6 ="; "   M7 = "; " "+
    "  M8 = "; "   M9 = "; "  M10 = "}.
compute mlab = {mlab; " M11 = "; " M12 = "; " M13 = "; " M14 = "; " M15 = "}.
compute mname = name(1,(2+nxd):(1+nxd+nm)).
do if (ncovs > 0).
   compute c = dd(:,(2+nxd+nm):(1+nxd+nm+ncovs)).
   compute x = {x, c}.
   compute xdata = x.
   compute cdata = c.
   compute xname = {xname, cname}.
end if.
compute nms={yname; t(mname); t(xname(1,1:(ncol(xname)-ncovs)))}.
do if (dumok = 1).
  compute nms={yname; t(mname); name(1,2)}.
end if.
do if (nm < 16).
  do if (dumok = 1).
    print dummat/title = "Coding of categorical X variable for analysis:"/cnames = xname2/format = 
    F5.2.
  end if.
end if.
compute boots=make((samples+1),((nx+1)*nm),-999).
do if (errs = 0 and criterr = 0).
  loop k = 1 to (samples+1).
    do if (k = 2 and savboot = 1).
      compute bootstrp = make(boot,(ncol(bootcoef)),-999).
    end if.
    compute bootcoef=0.
    do if (k > 1).
      compute u=trunc(uniform(n,1)*n)+1.
      compute xdata = x(u,:).
      compute mdata = m(u,:).
      compute ydata = y(u,:).
      do if (ncovs > 0).
       compute cdata = c(u,:).
      end if.
    end if.
    loop i = 1 to nm.
      compute start = 1+(i-1)*nx.
      compute xmat = {ones, xdata}.
      do if (covmy=2).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = mdata(:,i).
      compute amat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*amat.
      compute ssresid = csum((resid)&**2).       
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute r2cha = r2.
      compute adjr2cha = adjr2.
      do if (ncovs > 0 and covmy <> 2).
        compute xmat = {ones, cdata}.
        compute atmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute ssresid2 = csum((ymat-xmat*atmat)&**2).
        compute r2step1 = (sstotal-ssresid2)/sstotal.
        compute r2cha = r2-r2step1.
        compute adjr2cha = adjr2- (1-((1-r2step1)*(n-1)/(n-ncol(xmat)))).
        compute xmat = {ones, xdata}.
      end if.
      compute bootcoef={bootcoef,t(amat)}.
      compute r2b(i,1)=adjr2cha.
      do if (k = 1).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute tratio = amat&/sqrt(diag(covmat)).
        compute rnms = mname(1,i).
        print rnms/title = "*********************************************************************"+
    "*****"/rlabels = "Outcome:"/format = A8.
        compute lmat = ident(nrow(amat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*amat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*amat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = !decimals/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute op = {amat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute cov(start:(start+(nx-1)),start:(start+(nx-1)))=covmat(2:(ncol(covmat)-(ncovs*(covmy 
    <> 2))),2:(ncol(covmat)-(ncovs*(covmy <> 2)))).
        do if (covcoeff=1).
          print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
        end if.
      end if.
      compute mns(start:(start+(nx-1)),1)=amat(2:(1+nx)).
    end loop.
    compute xmat = {ones,mdata,xdata}.
    do if (covmy=1).
       compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
    end if.
    compute ymat = ydata.
      do if  (k = 1).
        print yname/title = "********************************************************************"+
    "******"/rlabels = "Outcome:"/format = A8.
        do if (ovals = 2).
          compute nmsd = {yname, "Analysis"}.
          print rcd/title = "Coding of binary DV for analysis:"/cnames = nmsd/format = F9.2.
        end if.
      end if.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute bootcoef={bootcoef,t(bmat)}.
        compute LL1 = LL2.
        compute covmat=varb.
      end if.
    do if (ovals <> 2).
      compute bmat =  inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
      compute resid=ymat-xmat*bmat.
      compute ssresid = csum((resid)&**2).
      compute r2 = (sstotal-ssresid)/sstotal.
      compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
      compute mse=ssresid/(n-ncol(xmat)).
      compute bootcoef={bootcoef,t(bmat)}.
    end if.
    do if (k = 1).
      do if (ovals = 2).
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms2 = {"Constant"; t(mname); t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms2/clabels = "Coeff" "se" "Z" 
    "p" "LLCI" "ULCI"/format=!decimals.
        compute opdirsum=outp((2+nmeds):(1+nmeds+nx),:).
        compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm))
    .
      end if.
      do if (ovals <> 2).
      hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
      compute lmat = ident(nrow(bmat)).
      compute lmat = lmat(:,2:ncol(lmat)).
      compute fratio = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/(ncol(xmat)-1).
      compute lmat = make(ncol(xmat),nx,0).
      compute lmat2 = ident(nx).
      compute lmat((2+nmeds):(2+nmeds+(nx-1)),:) = lmat2.
      compute fratiodr = (t(t(lmat)*bmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*bmat)))/nx.
      compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
      compute df2=(n-ncol(xmat)).
      compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), df2, pfr}. 
      print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", "df1", 
    "df2", "p".
      compute tratio = bmat&/sqrt(diag(covmat)).
      compute rnms2 = {"Constant"; t(mname); t(xname)}.
      compute op = {bmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat)))) }.
      compute temp=(n-ncol(xmat)).
      compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
      compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
      compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
      compute op={op,temp1,temp2}.
      print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms2/format = !decimals.
      compute opdirsum=op((2+nmeds):(1+nmeds+nx),:).
      compute cov((ncol(cov)-nm+1):ncol(cov),(ncol(cov)-nm+1):ncol(cov))=covmat(2:(1+nm),2:(1+nm)).
      compute xmat = {ones,mdata}.
      do if (ncovs > 0).
        compute xmat = {xmat, cdata}.
      end if.
      compute btmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
      compute ssresid2 = csum((ymat-xmat*btmat)&**2).
      compute r2direct = r2-((sstotal-ssresid2)/sstotal).
      compute pfr = 1-fcdf(fratiodr,(nx),(n-1-nx-nm-ncovs)).
      compute r2direct = {r2direct, fratiodr, nx, (n-1-nx-nm-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms2/cnames=rnms2/format=!decimals.
      end if.
    end if.
    do if (k = 1 and toteff = 1).
      print yname/title = "************************** TOTAL EFFECT MODEL "+
    "****************************"/rlabels = "Outcome:"/format = A8.
      compute xmat = {ones, xdata}.
      do if (covmy=1).
        compute xmat=xmat(:,1:(ncol(xmat)-ncovs)).
      end if.
      compute ymat = ydata.
      do if (ovals = 2).
        compute pt2 = make(n,1,(csum(ymat)/n)).
        compute LL3 = ymat&*ln(pt2)+(1-ymat)&*ln(1-pt2).
        compute LL3 = -2*csum(LL3).
        compute pt1 = make(n,1,0.5).
        compute bt1 = make(ncol(xmat),1,0).
        compute LL1 = 0.
        LOGIT2 pt1lp=pt1/xlp=xmat/ylp=ymat.
        do if (jjj > iterate).
          compute itprob = 2.
        end if.
        compute tval=xp2*xp2.
        compute bmat=b.
        compute LL1 = LL2.
        compute covmat=varb.
        compute LLdiff = LL3-LL2.
        compute pvalue=1-chicdf(LLdiff,(nrow(bmat)-1)).
        compute LL4 = LL2.
        compute mcF = LLdiff/LL3.
        compute cox = 1-exp(-LLdiff/n).
        compute nagel = cox/(1-exp(-(LL3)/n)).
        compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
        print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
        compute rnms = {"Constant"; t(xname)}.
        compute tstat = bmat&/sqrt(diag(covmat)).
        compute outp = {bmat,sqrt(diag(covmat)),tstat,(2*(1-cdfnorm(abs(tstat))))}.
        compute temp=bmat-abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        compute temp=bmat+abs(xp2)&*sqrt(diag(covmat)).
        compute outp = {outp, temp}.
        print outp/title = "Logistic Regression Model"/rnames = rnms/clabels = "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format=!decimals.
        compute optotsum=outp(2:(1+nx),:).
      end if.
      do if (ovals <> 2).
        compute cmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
        compute sstotal = t(ymat-(csum(ymat)/n))*(ymat-(csum(ymat)/n)).
        compute resid=ymat-xmat*cmat.
        compute ssresid = csum((resid)&**2). 
        compute r2 = (sstotal-ssresid)/sstotal.
        compute adjr2 = 1-((1-r2)*(n-1)/(n-ncol(xmat))).
        compute mse=ssresid/(n-ncol(xmat)).
        hcest hcdat=xmat/hcres=resid/hctog=hc3/hcmse=mse.
        compute lmat = ident(nrow(cmat)).
        compute lmat = lmat(:,2:ncol(lmat)).
        compute fratio = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/(ncol(xmat)-1).    
        compute pfr = 1-fcdf(fratio,(ncol(xmat)-1),(n-ncol(xmat))).
        compute op = {sqrt(r2), r2, mse, fratio, (ncol(xmat)-1), (n-ncol(xmat)), pfr}. 
        print op/title = "Model Summary"/format = F10.4/clabels = "R", "R-sq", "MSE", "F", 
    "df1", "df2", "p".
        compute tratio = cmat&/sqrt(diag(covmat)).
        compute op = {cmat, sqrt(diag(covmat)), tratio, (2*(1-tcdf(abs(tratio), n-ncol(xmat))))}.
        compute temp=(n-ncol(xmat)).
        compute temp =  (temp* (exp((temp-(5/6))*((xp2/(temp-(2/3)+(.11/temp)))*(xp2/(temp-(2/3)+
    (.11/temp)))))-1)).
        compute temp1 = op(:,1)-sqrt(abs(temp))&*op(:,2).
        compute temp2 = op(:,1)+sqrt(abs(temp))&*op(:,2).
        compute op={op,temp1,temp2}.
        compute rnms = {"Constant"; t(xname)}.
        print op/title = "Model"/clabels = "coeff", "se", "t", "p", "LLCI", "ULCI"/rnames = 
    rnms/format = !decimals.
        compute optotsum=op(2:(1+nx),:).
        compute ssresid2 = ssresid.
        compute r2total = r2.
        compute lmat = make(ncol(xmat),nx,0).
        compute lmat2 = ident(nx).
        compute lmat(2:(1+nx),:) = lmat2.
        compute fratio3 = (t(t(lmat)*cmat)*inv(t(lmat)*covmat*lmat)*((t(lmat)*cmat)))/nx.
        do if (ncovs > 0).
          compute xmat = {ones, cdata}.
          compute ctmat = inv(t(xmat)*xmat)*t(xmat)*ymat.
          compute ssresid2 = csum((ymat-xmat*ctmat)&**2).
          compute r2total = r2-((sstotal-ssresid2)/sstotal).
        end if.
        /* compute fratio3 = ((n-1-nx-ncovs)*r2total)/((1-r2)*nx) */. 
        compute pfr = 1-fcdf(fratio3,(nx),(n-1-nx-ncovs)).
        compute r2tot2 = {r2total, fratio3, nx, (n-1-nx-ncovs), pfr}.
      end if.
      do if (covcoeff=1).
        print covmat/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=rnms/cnames=rnms/format=!decimals.
      end if.
    end if.
    compute mns((nrow(mns)-nm+1):nrow(mns),1)=bmat(2:(1+nm),1).
    /* calculate relative indirect effects */.
    compute tmp = 1.
    loop i = 1 to (nm*nx).
      compute boots(k,i)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute boots(k,((nm*nx)+tmp))=r2b(tmp,1)*mns(((nm*nx)+tmp),1).
        compute tmp = tmp+1.
      end if.
    end loop.
    do if (savboot = 1 and k > 1).
      compute bootstrp((k-1),:)=bootcoef.
    end if.
  end loop.
  do if (method = 1).
    compute llci=make(1,ncol(boots),-999).
    compute ulci=make(1,ncol(boots),-999).
    compute indeff = t(boots(1,1:(nm*nx))).
    compute indeff22=t(boots(1,:)).
    compute omni = t(boots(1,((nm*nx)+1) : ncol(boots))).
    do if (nrow(boots) > 1).
      compute boots = boots(2:nrow(boots),:).      
      compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
      loop i = 1 to ncol(boots).
        bcboot databcbt = boots(:,i)/estmte=(indeff22(i,1)*bconoff)+(9999*(1-bconoff)).
        compute llci(1,i)=llcit.
        compute ulci(1,i)=ulcit.
        do if (badlo = 1 and llcit <> priorlo).
          compute badend={badend, llcit}.
          compute priorlo = llcit.
        end if.
        do if (badhi = 1 and ulcit <> priorhi).
          compute badend={badend, ulcit}.
          compute priorhi = ulcit.
        end if.
      end loop.
      compute llci = t(llci).
      compute ulci= t(ulci).
      compute llci={llci,ulci}.
    end if.
  end if.
  /* generate monte carlo samples */.
  do if (method = 2).
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute indeff(i,1)=mns(i,1)&*mns(((nm*nx)+tmp),1).
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = sqrt(-2*ln(uniform(reps,nrow(cov))))&*cos((2*3.14159265358979)*uniform(reps,
    nrow(cov))).
    compute boots=boots*chol(cov).
    loop i = 1 to ncol(boots).
      compute boots(:,i)=boots(:,i)+mns(i,1).
    end loop.
    compute tmp=1.
    loop i = 1 to (nm*nx).
      compute boots(:,i)=boots(:,i)&*boots(:,((nm*nx)+tmp)).
      compute temp = boots(:,i).
      compute temp(GRADE(boots(:,i))) = boots(:,i).
      compute boots(:,i) = temp.
      do if ((i/nx)=trunc(i/nx)).
        compute tmp = tmp+1.
      end if.
    end loop.
    compute boots = boots(:,1:(nm*nx)).
    compute bootse = t(sqrt(((reps*cssq(boots))-(csum(boots)&**2))/((reps-1)*reps))).
    compute llci = t({(boots(cilow,:));(boots(cihigh,:))}).
  end if.
  do if (toteff = 0).
    print/title = "******************** DIRECT AND INDIRECT EFFECTS *************************".
  else.
    print/title = "***************** TOTAL, DIRECT, AND INDIRECT EFFECTS ********************".
    compute rnmssum=rnms(2:(1+nx),1).
    do if (ovals <> 2).
      print optotsum/title = "Relative total effects of X of Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
      print r2tot2/title = "Omnibus test of total effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
    end if.
    do if (ovals = 2).
      print optotsum/title = "Relative total effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
    end if.
    print/title= "====="/space=0.
  end if.
  compute rnmssum=rnms2((2+nmeds):(1+nmeds+nx),1).
  do if (ovals <> 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff", "se", "t", 
    "p", "LLCI", "ULCI"/rnames = rnmssum/format = !decimals.
    print r2direct/title = "Omnibus test of direct effect of X on Y"/clabels = "R-sq", "F", 
    "df1", "df2", "p"/format = !decimals.
  end if.
  do if (ovals = 2).
    print opdirsum/title = "Relative direct effects of X on Y"/clabels = "coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/rnames = rnmssum/format = !decimals.
  end if.
  print/title= "====="/space=0.
  do if (omnibus = 1 and (method = 1 or method = 3)).
    compute indlbs = {indlbs; "Omnibus"}.
  end if.
  loop i = 1 to nm.
    compute op = mname(1,i).
    print op/title = "Relative indirect effect(s) of X on Y through:"/format = A8/space=0.
    do if (method = 2).
      compute clbs = {"Effect", "SE MC", "MC LLCI", "MC ULCI"}. 
      compute op = {indeff, bootse(1:(nm*nx)), llci}.
    end if.
    do if (method = 1).
      compute clbs = {"Effect", "SE(boot)", "LLCI", "ULCI"}.
      compute op = indeff.
      do if (samples > 0).
        compute op = {indeff, bootse(1:(nm*nx),:), llci(1:(nm*nx),:)}.
      end if.
    end if.
    compute temp = op((((i-1)*nx)+1):(nx*i),:).
    do if ((method = 1) and omnibus = 1).
      do if (samples = 0).
        compute temp={temp;omni(i,1)}.
      end if.
      do if (samples > 0).
        compute temp = {temp; omni(i,1), bootse(((nm*nx)+i),:), llci(((nm*nx)+i),:)}.
      end if.
    end if.
    print temp/title = " "/rnames = indlbs/cnames = clbs/format = !decimals/space=0.
    print/title = "----------".
  end loop.
end if.
do if (savboot = 1).
  compute bootstrp = bootstrp(:,2:ncol(bootstrp)).
  save bootstrp/outfile = *.
end if.
do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
end if.
do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
end if.
 END IF.
  /* end added  end of mediate */.
DO IF (mdpbe=1).
compute nm={yname,xname,mnames}.
do if (ncovs > 0).
  compute nm={yname,cnames,xname,mnames}.
end if.
compute tpx=xname. 
compute x = dd(:,2:ncol(dd)).  
compute tpy = yname.
compute tempy= dd(:,1).
compute n = nrow(dd).
compute nx=1.
compute criterr=0.
compute ddd = {"D1", "D2", "D3", "D4", "D5", "D6", "D7", "D8", "D9"}.
compute ddd1 = {"int_1", "int_2", "int_3", "int_4", "int_5", "int_6", "int_7", "int_8", "int_9"}.
compute dumok = 0.
compute mcfoc=mcx.
compute mcmod=mcm.
compute mcloc=(mcfoc>0).
compute con = make(n,1,1).
compute ncovs=ncol(x)-2.
do if (mcfoc > 0 or mcmod > 0).
  compute temp = dd.
  compute temp(GRADE(dd(:,(ncol(dd)-mcloc))),:) = dd.
  compute dd = temp.
  compute dummy = design(dd(:,ncol(dd)-mcloc)).
  compute nvls = ncol(dummy).
  compute nnvls = csum(dummy).
  compute toosmall=rsum(nnvls < 2).
  compute mnvls = cmin(t(nnvls)).
  do if ((rsum(nnvls < 2)) > 0).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=37.
  end if.
  do if (nvls > 10).
    compute criterr = 1.
    compute errs=errs+1.
    compute runerrs(errs,1)=36.
  end if.
  do if (criterr=0).
    compute dumok = 1.
    compute nnvls=make(nvls,1,0).
    compute nnvls(1,1)=dd(1,ncol(dd)-mcloc).
    compute temp = 2.
    loop i = 2 to n.
      do if (dd(i,ncol(dd)-mcloc) <> nnvls((temp-1),1)).
        compute nnvls(temp,1)=dd(i,ncol(dd)-mcloc).
        compute temp = temp+1.
      end if.
    end loop.
    compute dummy = dummy(:,2:ncol(dummy)).
    do if (mcfoc=4 or mcmod=4).
      compute minus1=make(1,ncol(dummy),-1).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) = 0).
          compute dummy(k,:)=minus1.
        end if.
      end loop.
    end if.
    do if (mcfoc = 2 or mcfoc = 3) or (mcmod = 2 or mcmod = 3)).
      loop k = 1 to n.
        do if (rsum(dummy(k,:)) > 0).
          loop i = 1 to ncol(dummy).
            do if (dummy(k,i) = 0).
              compute dummy(k,i) = 1.
            else.
              break.
            end if.
          end loop.
        end if.
      end loop.
      do if (mcfoc = 3 or mcmod=3).
        compute conmat1={-8,1,1,1,1,1,1,1,1;
                                        0,-7,1,1,1,1,1,1,1;
                                        0,0,-6,1,1,1,1,1,1;
                                        0,0,0,-5,1,1,1,1,1;
                                        0,0,0,0,-4,1,1,1,1;
                                        0,0,0,0,0,-3,1,1,1;
                                        0,0,0,0,0,0,-2,1,1;
                                        0,0,0,0,0,0,0,-1,1}.
        loop i = 1 to 8.
          compute conmat1(i,:)=conmat1(i,:)/(10-i).
        end loop.
        compute conmat1=t(conmat1((10-nvls):8,(10-nvls):9)).
        loop k=1 to n.
          compute dummy(k,:)=conmat1((rsum(dummy(k,:))+1),:). 
        end loop.
      end if.
    end if.
    compute nx = ncol(dummy).
    compute xname = ddd(1,1:nx).
    compute xdata = dummy.
    compute xname = ddd(1,1:nx).
    compute xname2={nm(1,(ncol(nm)-mcloc)), xname}.   
    compute indlbs = t(xname).
    compute dummat = make((nx+1),nx,0).
    compute dummat((2:nrow(dummat)),:)=ident(nx).
    do if (mcfoc = 2 or mcmod=2).
      loop i = 2 to nrow(dummat).
        loop j = 1 to (i-1).
          compute dummat(i,j) = 1.
        end loop.
      end loop.
    end if.
    do if (mcfoc = 3 or mcmod=3).
      compute dummat=conmat1.
    end if.
    do if (mcfoc = 4 or mcmod = 4).
       compute dummat(1,:)=minus1.
    end if.
    compute dummat={nnvls, dummat}.
  end if.
end if.
do if (criterr=0).
  compute y = dd(:,1).
  compute ovals = ncol(design(dd(:,1))).
  compute itprob = 0.
  do if (ovals = 2).
    compute omx = cmax(y(:,1)).
    compute omn = cmin(y(:,1)).
    compute y(:,1) = (y(:,1) = omx).
    compute rcd = {omn, 0; omx, 1}.
  end if.
  compute sstotal = csum((y-(csum(y)/n))&**2).
  compute outv = t(nm(1,1)).
  compute mdtr = nm(1,ncol(dd)).
  compute fciv = nm(1,(ncol(dd)-1)).
  do if (mcfoc <> 0 or mcmod <> 0).
    compute inter=make(n,nx,0).
    loop i = 1 to nx.
      compute inter(:,i)=dummy(:,i)&*dd(:,ncol(dd)-(1-mcloc)).
    end loop.
    do if (ncovs = 0).
      compute x={con,dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={nm(1,ncol(dd)-(1-mcloc)); t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
    do if (ncovs > 0).
      compute x={con,dd(:,2:(1+ncovs)),dd(:,ncol(dd)-(1-mcloc)),dummy}.
      compute x={x,inter}.
      compute nms={t(nm(1,2:(1+ncovs))); nm(1,ncol(dd)-(1-mcloc)); 
    t(ddd(1,1:ncol(dummy)));t(ddd1(1,1:ncol(dummy)))}.
    end if.
  end if.
  compute xmns = csum(x)&/n.
  compute focvals = {1,0,0}.
  compute highwarn = 0.
  compute lowwarn = 0.
  do if (ncol(x) > (2*nx+2)).
    compute covmns = {1, xmns(1,2:(ncol(x)-1-(2*nx)))}.
    compute focvals = {covmns,0,0}.
  end if.
  compute dfres = n-ncol(x).
  do if (ovals = 2).
    compute pt2 = make(n,1,(csum(y)/n)).
    compute LL3 = y&*ln(pt2)+(1-y)&*ln(1-pt2).
    compute LL3 = -2*csum(LL3).
    compute pt1 = make(n,1,0.5).
    compute bt1 = make(ncol(x),1,0).
    compute LL1 = 0.
    LOGIT2 pt1lp=pt1/xlp=x/ylp=y.
    do if (jjj > iterate).
      compute itprob = 2.
    end if.
    compute tval=xp2*xp2.
  else if (ovals <> 2).
    compute invXtX=inv(t(x)*x).
    compute b = invXtX*t(x)*y.
    compute resid = y-(x*b).
    compute ssresid=csum(resid&*resid).
    compute msresid=ssresid/dfres.
    compute varb = msresid*invXtX.
    compute seint=sqrt(varb(nrow(varb),nrow(varb))).
    compute k3 = nrow(b).
    do if (hc3 = 1).
      compute xhc=x.
      compute h = xhc(:,1).
      loop i3=1 to n.
        compute h(i3,1)= xhc(i3,:)*invXtX*t(xhc(i3,:)). 
      end loop.
      loop i3=1 to k3.
        compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
      end loop.
      compute varb=(invXtX*t(xhc)*xhc*invXtX).
    end if.
    compute seb = sqrt(diag(varb)).
    compute r2 = 1-(ssresid/sstotal).
    do if (mcfoc <> 0 or mcmod <> 0).
      compute ytp=y.
      compute xtp=x(:,1:(ncol(x)-nx)).
      compute btp = inv(t(xtp)*xtp)*t(xtp)*ytp.
      compute residt = ytp-(xtp*btp).
      compute ssresidt=csum(residt&*residt).
      compute r2noint = 1-(ssresidt/sstotal).
    end if.
    compute pr = ncol(x)-1.
    compute lmat = ident(nrow(b)).
    compute lmat = lmat(:,2:ncol(lmat)).
    compute f = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/pr).
    compute pf = 1-fcdf(f,pr,dfres).
    compute pf = {sqrt(r2),r2,msresid,f,pr,dfres,pf}.
    compute xd=abs(xp2).
  end if.
  do if (itprob = 0).
    compute tstat = b&/seb.
    do if (ovals <> 2).
      compute p = 2*(1-tcdf(abs(tstat), dfres)).
    else if (ovals = 2).
      compute p = 2*(1-cdfnorm(abs(tstat))).
    end if.
    compute outp = {b,seb,tstat,p}.
    compute nms = {"constant"; nms; "interact"}.
    do if (ovals <> 2).
      compute tval =  (dfres* (exp((dfres-(5/6))*((xd/(dfres-(2/3)+(.11/dfres)))*(xd/(dfres-(2/3)+
    (.11/dfres)))))-1)).
      compute outp={outp,(b-sqrt(tval)&*seb),(b+sqrt(tval)&*seb)}.
    end if.
    compute bb = tval.
    do if (dumok = 1).
      do if (mcfoc > 0).
        print dummat/title = "Coding of categorical X variable for analysis:"/cnames = 
    xname2/format = F5.2.
      else if (mcmod > 0).
        print dummat/title = "Coding of categorical M variable for analysis:"/cnames = 
    xname2/format = F5.2.
      end if.
    end if.
    print yname/title = "**************************************************************************"+
    ""/rlabels = "Outcome:"/format = A8.
    do if (ovals = 2).
      compute nmsd = {outv, "Analysis"}.
      print rcd/title = "Coding of binary Y for analysis:"/cnames = nmsd/format = !decimals.
      compute outp = {outp,(outp(:,1)-sqrt(tval)&*outp(:,2)),(outp(:,1)+sqrt(tval)&*outp(:,2))}.
      compute LLdiff = LL3-LL2.
      compute pvalue=1-chicdf(LLdiff,(nrow(b)-1)).
      compute LL4 = LL2.
      compute mcF = LLdiff/LL3.
      compute cox = 1-exp(-LLdiff/n).
      compute nagel = cox/(1-exp(-(LL3)/n)).
      compute pf = {LL2, LLdiff, pvalue, mcF, cox, nagel, n}.
      print pf/title = "Logistic Regression Summary"/clabels = "-2LL" "Model LL" "p-value" 
    "McFadden" "CoxSnell" "Nagelkrk" "n"/format !decimals.
      print outp/title = "Logistic Regression Model"/rnames = nms/clabels "Coeff" "se" "Z" "p" 
    "LLCI" "ULCI"/format !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      compute varbtmp=varb.
      compute btmp=b.
      compute LL2f=LL2.
      compute pt1lp = make(n,1,0.5).
      compute ylp=y.
      compute xlp=x(:,1:(ncol(x)-nx)).
      compute bt1 = make(ncol(xlp),1,0).
      LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
      compute LLdiff=LL2-LL2f.
      compute b=btmp.
      compute varb=varbtmp.
      compute pchi=1-chicdf(lldiff,nx).
      compute rcha={LLdiff,nx,pchi}.
    end if.
    do if (ovals <> 2).
      print pf/title = "Model Summary"/clabels = "R" "R-sq" "MSE" "F" "df1" "df2" "p"/format 
    !decimals.
      compute lmat=make(nrow(b),1,0).
      compute lmat(nrow(lmat),1)=1.
      compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/1).
      compute rcha=((b(nrow(b),1)/seint)*(b(nrow(b),1)/seint))*(1-r2)/dfres.
      compute rcha = {rcha, fcha, 1, dfres, outp((pr+1),4)}.
      print outp/title = "Model"/rnames = nms/clabels "coeff" "se" "t" "p" "LLCI" "ULCI"/format 
    !decimals.
      do if (covcoeff=1).
        print varb/title="Covariance matrix of regression parameter "+
    "estimates"/rnames=nms/cnames=nms/format=!decimals.
      end if.
      do if (mcfoc > 0 or mcmod > 0).
        compute rcha=r2-r2noint.
        compute lmat=make((nrow(b)-nx),nx,0).
        compute lmat2=ident(nx).
        compute lmat={lmat;lmat2}.
        compute fcha = (t(t(lmat)*b)*inv(t(lmat)*varb*lmat)*((t(lmat)*b)))/nx).
        compute pvalr2c=1-fcdf(fcha,nx,dfres).
        compute rcha = {rcha, fcha, nx, dfres,pvalr2c}.
      end if.
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute intkey = {"a", "b", "c", "d", "e"}.
      loop i = 1 to nx.
        compute intkey={intkey; ddd1(1,i), " : ", ddd(1,i), " X ", nm(1,ncol(dd)-(1-mcloc))}.
      end loop.
      compute intkey=intkey(2:nrow(intkey),:).
      print intkey/title="Product terms key:"/format = a8.
    end if.
    do if (ovals <> 2).
      print rcha/title = "R-square increase due to interaction:"/clabels "R2-chng" "F" "df1" 
    "df2" "p"/format !decimals.
    else if (ovals = 2).
      print rcha/title = "Likelihood ratio test for interaction:"/clabels "Chi-sq" "df" "p"/format 
    !decimals.
    end if.
    print/title = "***************************************************************************".
    compute mdvar = x(:,(ncol(x)-1)).
    do if (mcfoc = 0 and mcmod = 0).
      compute g1 = b((ncol(x)-2),1).
      compute g3 = b(ncol(x),1).
      compute vg1 = varb((ncol(x)-2),(ncol(x)-2)).
      compute vg3 = varb(ncol(x),ncol(x)).
      compute covg1g3 = varb((ncol(x)-2), ncol(x)).
    end if.
    do if (mcfoc > 0 or mcmod > 0).
      compute mdvar = x(:,(ncol(x)-(2*nx))).
    end if.
    compute mdmin = cmin(mdvar).
    compute mdmax = cmax(mdvar).
    compute fvar = x(:,(ncol(x)-2)).
    compute nval = ncol(design(mdvar)).
    compute fvmin = cmin(fvar).
    compute fvmax = cmax(fvar).
    do if (mmodval = 999).
      compute mnmd = csum(mdvar)/n.
      compute tmp = make(n,1,mnmd).
      compute sdmd = sqrt(csum(((mdvar-tmp)&**2))/(n-1)).
      compute probeval = {mnmd-sdmd; mnmd; mnmd+sdmd}.
      do if (probeval(1,1) < mdmin).
        compute lowwarn = 1.
        compute probeval(1,1)=mdmin.
      end if.
      do if (probeval(nrow(probeval),1) > mdmax).
        compute highwarn = 1.
        compute probeval(nrow(probeval),1)=mdmax.
      end if.
      do if (quantile = 1).
        compute tmp = mdvar.
        compute tmp(GRADE(mdvar(:,1)),:) = mdvar.
        compute mdvar = tmp.
        compute probeval={mdvar(trunc(0.1*n),1);mdvar(trunc(0.25*n),1);mdvar(trunc(0.5*n),1)
    ;mdvar(trunc(0.75*n),1);mdvar(trunc(0.9*n),1)}.
      end if.
    end if.
    do if (nval = 2).
      compute probeval = make(2,1,0).
      /* compute probeval(1,1)=mdvar(1,1) */.
      compute probeval(1,1)=cmin(mdvar).
      loop i = 1 to n.
        do if (mdvar(i,1) <> probeval(1,1)).
          compute probeval(2,1) = mdvar(i,1).
          BREAK.
        end if.
      end loop.
    end if.
    do if (mmodval <> 999).
      compute probeval = mmodval.
    end if.
    compute outp = make(nrow(probeval),7,0).
    do if (mcfoc > 0 or mcmod > 0).
      compute focvals=make(1,ncol(x)+1,1).
      do if (mcfoc > 0).
        print/title = "Conditional effect of X on Y at values of the moderator:".
        print/title = " "/space=0.
        compute rnn2=mdtr.
        compute matt=make(nx,6,0).
      end if.
      loop jj=1 to nrow(probeval).
        do if (mcfoc > 0).
          loop ii=1 to nx.
            compute g1=b((ncol(x)-(2*nx)+ii),1).
            compute g3=b((ncol(x)-nx+ii),1).
            compute vg1=varb((ncol(x)-(2*nx)+ii),(ncol(x)-(2*nx)+ii)).
            compute vg3=varb((ncol(x)-nx+ii),(ncol(x)-nx+ii)).
            compute covg1g3=varb((ncol(x)-(2*nx)+ii),(ncol(x)-nx+ii)).
            compute x2 = probeval(jj,1).
            compute w1 = g1+g3*x2.
            compute varw1 = vg1+(2*x2*covg1g3)+((x2*x2)*vg3).
            compute sew1 = sqrt(varw1).
            compute t1 = w1/sew1.
            do if (ovals <> 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-tcdf(abs(t1), dfres)).
              compute cnms = {"Coeff", "se", "t", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            else if (ovals = 2).
              compute LLCI = (w1-sqrt(tval)&*sew1).
              compute ULCI = (w1+sqrt(tval)&*sew1).
              compute p = 2*(1-cdfnorm(abs(t1))).        
              compute cnms = {"Coeff", "se", "Z", "p", "LLCI", "ULCI"}.
              compute matt(ii,:)={w1,sew1,t1,p,llci,ulci}.
            end if.
          end loop.
          compute rnms=t(xname).
          compute mdvalpr=probeval(jj,1).
          print mdvalpr/title = "Moderator value:"/rnames=rnn2/format=!decimals/space=0.
          print matt/title=" "/cnames=cnms/rnames=rnms/format=!decimals/space=0.
          compute xprob=x(:,(ncol(x)-(2*nx)))-mdvalpr.
          loop kk = 1 to nx.
            compute xprob={xprob, (xprob(:,1)&*x(:,(ncol(x)-(2*nx)+kk)))}.          
          end loop.
          compute xprob={x(:,1:((ncol(x)-(2*nx))-1)),xprob}.
          do if (ovals <> 2).
            compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
            compute residc = y-(xprob*bmultc).
            compute ssresidc=csum(residc&*residc).
            compute r2c = r2-(1-(ssresidc/sstotal)).
            compute fcha2=(dfres*r2c)/(nx*(1-r2)).
            do if (hc3 = 1).
              loop kk = 1 to nx.
                compute xprob={xprob, x(:,(ncol(x)-(2*nx)+kk))}.          
              end loop.
              compute bmultc = inv(t(xprob)*xprob)*t(xprob)*y.
              compute k3 = nrow(bmultc).
              compute xhc=xprob.
              compute h = xhc(:,1).
              loop i3=1 to n.
                compute h(i3,1)= xhc(i3,:)*inv(t(xprob)*xprob)*t(xhc(i3,:)). 
              end loop.
              loop i3=1 to k3.
                compute xhc(:,i3) = (resid(:,ncol(resid))&/(1-h))&*xhc(:,i3).
              end loop.
              compute varbc=(inv(t(xprob)*xprob)*t(xhc)*xhc*inv(t(xprob)*xprob)).
              compute lmat=make((nrow(bmultc)-nx),nx,0).
              compute lmat2=ident(nx).
              compute lmat={lmat;lmat2}.
              compute fcha2 = (t(t(lmat)*bmultc)*inv(t(lmat)*varbc*lmat)*((t(lmat)*bmultc)))/nx).
            end if.
            compute pvalr2cc=1-fcdf(fcha2,nx,dfres).
            compute rcha2 = {r2c, fcha2, nx, dfres,pvalr2cc}.
            print rcha2/title = "Test of equality of conditional means at this value of the "+
    "moderator"/clabels "R2-chng" "F" "df1" "df2" "p"/format !decimals.
          else if (ovals = 2).
            compute btmp=b.
            compute pt1lp = make(n,1,0.5).
            compute ylp=y.
            compute xlp=xprob.
            compute bt1 = make(ncol(xlp),1,0).
            LOGIT2 pt1lp=pt1lp/xlp=xlp/ylp=ylp.
            compute LLdiff=LL2-LL4.
            compute b=btmp.
            compute varb=varbtmp.
            compute pchi=1-chicdf(lldiff,nx).
            compute rcha2 = {LLdiff, nx,pchi}.
            print rcha2/title = "Test of equality of log odds conditioned on this moderator "+
    "value"/clabels "Chi-sq" "df" "p"/format !decimals.
          end if.
        end if.
        /* end of do if mcfoc > 0 */.
        compute ttttt=make(nrow(dummat),1,probeval(jj,1)).
        compute ttttt={ttttt,dummat(:,2:ncol(dummat))}.
        loop kkk=1 to nx.
          compute ttttt={ttttt,ttttt(:,1)&*ttttt(:,1+kkk)}.
        end loop.
        compute ones=make(nrow(dummat),1,1).
        do if (ncol(x) > (2*nx+2)).
          compute covmnmat=make(nrow(ttttt),ncol(covmns),0).
          loop kkk=1 to nrow(ttttt).
            compute covmnmat(kkk,:)=covmns.  
          end loop.
          compute ttttt={covmnmat,ttttt}.
        else.
          compute ttttt={ones,ttttt}.
        end if.
        compute focvals={focvals;dummat(:,1),ttttt}.  
        do if (mcfoc > 0).
          compute yhat={dummat(:,1),(ttttt*b)}.
          compute cmnms={fciv, "yhat"}.
          do if (ovals <> 2).
            print yhat/title = "Estimated conditional means at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (ovals = 2).
            print yhat/title = "Estimated conditional log odds at this value of the "+
    "moderator"/cnames=cmnms/format !decimals.
          end if.
          do if (jj <> nrow(probeval)).
            print/title="-------------"/space=0.
          end if.
        end if.
      end loop.
      compute focvals=focvals(2:nrow(focvals),:).
      compute yhat=focvals(:,2:ncol(focvals))*b.
      compute focvals={focvals(:,1),focvals(:,(ncol(focvals)-(2*nx))),yhat}.
      compute cnms={fciv,mdtr,"yhat"}.
      do if (mcmod > 0).
        compute cnms={mdtr,fciv,"yhat"}.
      end if.
    end if.
    do if (mcmod > 0).
      compute outp=make((nx+1),7,0).
      compute bcatm={b((ncol(x)-(2*nx)),1);b((ncol(x)-nx+1):ncol(x),1)}.
      compute outp(:,1)=dummat(:,1).
      compute bcatcov=varb((ncol(x)-nx):ncol(x),(ncol(x)-nx):ncol(x)).
      compute bcatcov(1,1)=varb((ncol(x)-(2*nx)),(ncol(x)-(2*nx))).
      compute bcatcov(2:nrow(bcatcov),1)=varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx)).
      compute bcatcov(1,2:nrow(bcatcov))=t(varb((ncol(x)-nx+1):ncol(x),(ncol(x)-2*nx))).
      loop i = 1 to nrow(dummat).
        compute catmval={1,dummat(i,2:ncol(dummat))}.
        compute condeff=catmval*bcatm.
        compute condse=sqrt(catmval*bcatcov*t(catmval)).
        compute outp(i,2:3)={condeff,condse}.
      end loop.
      compute outp(:,4)=outp(:,2)&/outp(:,3).
      compute outp(:,5) = 2*(1-tcdf(abs(outp(:,4)), dfres)).
      compute outp(:,6) = (outp(:,2)-sqrt(tval)&*outp(:,3)).
      compute outp(:,7) = (outp(:,2)+sqrt(tval)&*outp(:,3)).
      do if (ovals <> 2).
        compute cnmms = {xname2(1,1), "coeff", "se", "t", "p", "LLCI", "ULCI"}.
      end if.
      do if (ovals = 2).
        compute outp(:,5) = 2*(1-cdfnorm(abs(outp(:,4)))).
        compute cnmms = {xname2(1,1), "coeff", "se", "Z", "p", "LLCI", "ULCI"}.
      end if.
      print outp/title = "Conditional Effect of Focal Predictor in Groups Defined by the "+
    "Moderator Variable:"/cnames=cnmms/format = !decimals.
    end if.
    do if (nval > 2 and (mmodval = 999) and mcmod = 0).
      do if (quantile <> 1).
        print/title = "Moderator values are the sample mean and plus/minus one SD from mean".
      else.
        print/title = "Values for quantitative moderators are 10th, 25th, 50th, 75th, and 90th "+
    "percentiles".
      end if.
      do if (highwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD above the mean was".
         print/title = "replaced with the maximum because one SD above the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
      do if (lowwarn = 1 and quantile = 0).
         print/title = "NOTE: For the moderator values above, one SD below the mean was".
         print/title = "replaced with the minimum because one SD below the mean is outside"/space=0.    
        print/title = "of the range of the data."/space=0.  
      end if.
    end if.
    do if (nval = 2 and (mmodval = 999)).
      print/title = "The moderator variable is dichotomous".
    end if.
    do if (mmodval <> 999 and mcmod = 0 and ((mmodval < mdmin) or (mmodval > mdmax))).
      print/title = "Warning: Moderator value specified is outside of the range of the data".
    end if.
    do if (plot = 1).
      do if (mcfoc = 0 and mcmod=0).
        compute cnms = {t(nms((ncol(x)-2):(ncol(x)-1),1)), "yhat"}.
      end if.
      do if (ovals = 2).
        compute prob = exp(focvals(:,3))&/(1+exp(focvals(:,3))).
        compute focvals = {focvals, prob}.
        compute cnms = {cnms, "prob"}.
      end if.
      print/title = "***************************************************************************".
      !let !line0 = !concat("DATA LIST FREE/", !m, " ", !x, " ", !y, ".").
      !let !line01 = !concat("DATA LIST FREE/", !x, " ", !m, " ", !y, ".").
      !let !line02 = !concat("DATA LIST FREE/", !m, " ", !x, " ", " logodds prob.").
      !let !line03 = !concat("DATA LIST FREE/", !x, " ", !m, " ", " logodds prob.").
      !let !line1 = "BEGIN DATA.".
      !let !line2 = "END DATA.".
      !let !line3 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH ", !y, " BY ", !m, ".").
      !let !line31 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH ", !y, " BY ", !x, ".").
      !let !line32 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH logodds BY ", !m, ".").
      !let !line33 = !concat("GRAPH/SCATTERPLOT=", !x, " WITH prob BY ", !m, ".").
      !let !line34 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH logodds BY ", !x, ".").
      !let !line35 = !concat("GRAPH/SCATTERPLOT=", !m, " WITH prob BY ", !x, ".").
      print/title = "Data for visualizing conditional effect of X on Y".
      print/title = "Paste text below into a SPSS syntax window and execute to produce "+
    "plot."/space=0.
      do if (ovals = 2).
        do if (mcfoc=0).
          print/title = !quote(!line02)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line03)/space=1.
        end if.
      else if (ovals <> 2).
        do if (mcfoc=0).
          print/title = !quote(!line0)/space=1.
        end if.
        do if (mcmod=0).
          print/title = !quote(!line01)/space=1.
        end if.
      end if.
      print/title = !quote(!line1)/space=0.
      print focvals/title = " "/format = !decimals/space=0.
      print/title = !quote(!line2)/space=1.
      do if (ovals <> 2 and mcmod=0).
        print/title = !quote(!line31)/space=0.
      end if.
      do if (ovals <> 2 and mcfoc=0).
        print/title = !quote(!line3)/space=0.
      end if.
      do if (ovals = 2 and mcmod=0).
        print/title = !quote(!line34)/space=0.
        print/title = !quote(!line35)/space=0.
      end if.
      do if (ovals = 2 and mcfoc=0).
        print/title = !quote(!line32)/space=0.
        print/title = !quote(!line33)/space=0.
      end if.
      do if (ncovs > 0).
        print/title = "NOTE: For data above, covariates are set to their sample means.".
      end if.
    end if.
  end if.
  do if (itprob = 1).
    print/title = "ERROR: There was a problem during iteration.".
  end if.
  do if (itprob = 2).
    compute errs = errs+1.
    compute runerrs(errs,1) = 22.
  end if.
end if. 
  END IF.
end if.
do if (bad > 0).
  compute note(notes,1) = 9.
  compute notes = notes + 1.
end if.
print/title = "******************** ANALYSIS NOTES AND WARNINGS *************************".
loop i = 1 to errs.
  do if (runerrs(i,1) = 1).
    print/title = "ERROR: One of your declared mediators is dichotomous.  This procedure can't "+
    "be used.".
  end if.
  do if (runerrs(i,1) = 2).
    print/title = "ERROR: For model 6, this procedure limits the number of mediators to four.".
  end if.
  do if (runerrs(i,1) = 3).
    print/title = "ERROR: For models 1, 2, and 3, only a single variable can be listed in the M "+
    "list.".
  end if. 
  do if (runerrs(i,1) = 4).
    print/title = "ERROR: You requested a model involving W but did not provide a valid W "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 5).
    print/title = "ERROR: You requested a model involving Z but did not provide a valid Z "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 6).
    print/title = "ERROR: You requested a model involving Q but did not provide a valid Q "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 7).
    print/title = "ERROR: You requested a model involving V but did not provide a valid V "+
    "variable name.".
  end if. 
  do if (runerrs(i,1) = 8).
    print/title = "ERROR: You specified a W variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 9).
    print/title = "ERROR: You specified a Z variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 10).
    print/title = "ERROR: You specified a Q variable for a model that does not need it".
  end if. 
  do if (runerrs(i,1) = 11).
    print/title = "ERROR: You specified a V variable for a model that does not need it.".
  end if. 
  do if (runerrs(i,1) = 12).
    print/title = "ERROR: The variable specified for W has already been assigned.".
  end if.
  do if (runerrs(i,1) = 13).
    print/title = "ERROR: The variable specified for Z has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 14).
    print/title = "ERROR: The variable specified for Q has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 15).
    print/title = "ERROR: The variable specified for V has already been assigned.".
  end if.   
  do if (runerrs(i,1) = 16).
    print/title = "ERROR: You did not provide a valid Y variable name.".
  end if. 
  do if (runerrs(i,1) = 17).
    print/title = "ERROR: The variable specified for Y has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 18).
    print/title = "ERROR: Model 6 requires more than one mediator.".
  end if. 
  do if (runerrs(i,1) = 19).
    print/title = "ERROR: You have not specified a valid model number.".
  end if. 
  do if (runerrs(i,1) = 20).
    print/title = "ERROR: At least one and only one variable must be listed for X.".
  end if.
  do if (runerrs(i,1) = 21).
    print/title = "ERROR: At least one and only one variable must be listed for Y.".
  end if. 
  do if (runerrs(i,1) = 22).
    print/title = "ERROR: Iteration didn't converge to a solution.  Interpret results with "+
    "caution.".
  end if. 
  do if (runerrs(i,1) = 23).
    print/title = "ERROR: You specified a clustering variable that does not exist in your "+
    "variable list.".
  end if. 
  do if (runerrs(i,1) = 24).
    print/title = "ERROR: You specified a clustering variable that has already been assigned.".
  end if. 
  do if (runerrs(i,1) = 25).
    print/title = "ERROR: One or more of your M variables is not listed in the variables list.".
  end if.
  do if (runerrs(i,1) = 26).
    print/title = "ERROR: A maximum of 20 cluster units is allowed.  Use multilevel modeling "+
    "instead.".
  end if.
  do if (runerrs(i,1) = 27).
    print/title = "ERROR: One of the variables in your model is a constant.".
  end if.
  do if (runerrs(i,1) = 28).
    print/title = "ERROR: Dichotomous Y is not permitted with WS option.".
  end if.
  do if (runerrs(i,1) = 29).
    print/title = "ERROR: Insufficient number of variables in vars= list when using WS option.".
  end if.
  do if (runerrs(i,1) = 30).
    print/title = "ERROR: Too many variables in vars= list when using WS option.  Covariates not "+
    "allowed.".
  end if.
  do if (runerrs(i,1) = 31).
    print/title = "ERROR: mmodval and wmodval can't both be set to zero with contrast option.".
  end if.
  do if (runerrs(i,1) = 32).
    print/title = "ERROR: You did not provide a valid X variable name.".
  end if.
  do if (runerrs(i,1) = 33).
    print/title = "ERROR: PROCESS requires all variable names to be eight characters or fewer.".
    print/title = "       Please shorten variable names and reexecute."/space=0.
  end if.
  do if (runerrs(i,1)=34).
    print/title = "ERROR: X and M cannot both be specified as multicategorical.".
  end if.
  do if (runerrs(i,1)=35).
    print/title = "ERROR: Multicategorical variable specification available only for models 1 "+
    "and 4.".
  end if.
  do if (runerrs(i,1)=36).
    print/title = "ERROR: Categorical variables cannot have more than 10 categories.".
  end if.
  do if (runerrs(i,1)=37).
    print/title = "ERROR: Each group must have at least two cases.".
  end if.
  do if (runerrs(i,1)=38).
    print/title = "ERROR: Cluster option not available with multicategorial variables.".
  end if.
  do if (runerrs(i,1)=39).
    print/title = "ERROR: A maximum of 10 variables is allowed in the M= list.".
  end if.
  do if (runerrs(i,1) = 41).
    print/title = "ERROR: MCM option not permitted in model 4.".
  end if.
end loop.
do if (errs = 0).
  do if (boot > 1 or mc > 0).
    do if (bconoff = 1 and boot > 1).
      print boot/title = "Number of bootstrap samples for bias corrected bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (bconoff = 0 and boot > 1).
      print boot/title = "Number of bootstrap samples for percentile bootstrap confidence "+
    "intervals:"/format = F8.0.
    end if.
    do if (mc > 1).
      print mc/title = "Number of samples for Monte Carlo confidence intervals:"/format = F8.0.
    end if.
    do if (booterr = 1).
      compute badend = badend(1,2:ncol(badend)).
      print badend/title = "WARNING: Bootstrap CI endpoints below not trustworthy.  Decrease "+
    "confidence or increase bootstraps"/format = !decimals.
    end if.
  end if.
  print conf/title = "Level of confidence for all confidence intervals in output:"/format = F8.2.
  do if ((center = 1 or ws=1) and (ncol(centvar) > 1)).
    compute centvar = centvar(1,2:ncol(centvar)).
    print centvar/title = "NOTE: The following variables were mean centered prior to "+
    "analysis:"/format = a8.
  end if.
  loop i = 1 to notes.
    do if (note(i,1) = 1).
      print/title = "NOTE: Confidence level restricted to between 50 and 99.9999%.  95% "+
    "confidence is provided in output".
    end if.
    do if (note(i,1) = 2).
      print/title = "NOTE: Effect size measures for indirect effects are not available for "+
    "models with dichotomous outcomes".
    end if.
    do if (note(i,1) = 3).
      print/title = "NOTE: All standard errors for continuous outcome models are based on the "+
    "HC3 estimator".
    end if.
    do if (note(i,1) = 6).
      print/title = "NOTE: The number of bootstrap samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 7).
      print/title = "NOTE: The Johnson-Neyman method is available only for Models 1 and 3".
    end if.
    do if (note(i,1) = 8).
      print/title = "NOTE: The Johnson-Neyman method cannot be used with a dichotomous moderator".
    end if.
    do if (note(i,1)=9).
      print bad/title = "NOTE: Some bootstrap samples had to be replaced.  The number of such "+
    "replacements was:".
    end if.
    do if (note(i,1) = 11).
      print nmiss/title = "NOTE: Some cases were deleted due to missing data.  The number of "+
    "such cases was:".
    end if.
    do if (note(i,1) = 12).
      print /title = "NOTE: Monte Carlo method available only for models 4 and 5.  Bootstrapping "+
    "was used instead.".
    end if.
    do if (note(i,1) = 13).
      print/title = "NOTE: The number of Monte Carlo samples was adjusted upward given your "+
    "desired confidence". 
    end if.
    do if (note(i,1) = 19).
      print/title = "NOTE: Effect sizes not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 16).
      print/title = "NOTE: Normal theory tests not available for within-subject analyses.". 
    end if.
    do if (note(i,1) = 17).
      print/title = "NOTE: Monte Carlo confidence intervals not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 18 and warnrep=0).
      print/title = "WARNING: You have requested OLS estimation with a dichotomous criterion.".
      print/title = "Interpret model coefficients and inferential statistics with "+
    "caution."/space=0. 
      compute warnrep=1.
    end if.
    do if (note(i,1) = 20).
      print/title = "NOTE: Saving of bootstrap estimates not available for within-subject "+
    "analyses.". 
    end if.
    do if (note(i,1) = 22).
      print/title = "NOTE: Effect size option with covariates requires covariates in models of M "+
    "and Y.".
    end if.
    do if (note(i,1) = 23).
      print/title = "NOTE: Johnson-Neyman method not available with multicategorical X".
    end if.
    do if (note(i,1) = 24).
      print/title = "NOTE: MMODVAL option not available with a multicategorical moderator".
    end if.
    do if (note(i,1) = 25).
      print/title = "NOTE: Omnibus test for relative indirect effects is not available with the "+
    "MC option.".
    end if.
    do if (note(i,1) = 26).
      print/title = "NOTE: CONTRAST option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 27).
      print/title = "NOTE: EFFSIZE option not available with multicategorical X.".
    end if.
    do if (note(i,1) = 28).
      print/title = "NOTE: NORMAL option not available with multicategorical X.".
    end if.
  end loop.
end if.
end matrix.
match files/file = */drop = w999999t z999999t q999999t v999999t.
execute.
set printback = on.
!ENDDEFINE.
/* DO NOT ATTEMPT TO MODIFY OR EXECUTE THE CODE BELOW */.
/* DIALOG BOX USERS SHOULD CLICK "OK" TO EXECUTE PROCESS */.
/* DO NOT USE THE PASTE BUTTON IN THE DIALOG BOX */.
PRCD123 vars = ean coplay xp 
  /y=ean/x=coplay/m=xp/w=/z=/v=/q=/
  model =4/boot=1000/center=0/hc3=0/effsize=0/
  normal=0/coeffci=1/conf=95/percent=0/total=0/
  covmy=0/jn=0/quantile =0/plot=0/contrast=0/
  decimals=F10.4/covcoeff=0 
  .
restore.