Maximum likelihood moderated regression with incomplete predictors and interaction

DATA:

file = conscientiousness.dat;

VARIABLE:

names = agree consc jobperf;

usevariables = x z xz y;

missing = all (-99);

DEFINE:

! define X as conscientiousness;

x = consc; 

! define Z as agreeableness;   

z = agree;

! compute product term; 

xz = x*z; 

! define Y as job performance;    

y = jobperf;  

ANALYSIS: 

estimator = mlr;

MODEL:

! fix direct loadings to 1;

latentx by x@1;               

latentz by z@1;

latentxz by xz@1;

! fix predictor variable residual variances to 0;

x@0; z@0; xz@0;    

! Cross-loadings;

latentx by xz (latxbyxz);      

latentz by xz (latzbyxz);

! predictor variable means (measurement intercepts);

[x] (xmean);                  

[z] (zmean);

[xz];

! latent variable means;

[latentx] (lxmean) ;                 

[latentz] (lzmean);

[latentxz] (lxzmean);

! latent variable variances;

latentx latentz latentxz;

! latent variable covariances;

latentx with latentz (covlxlz);     

latentx with latentxz;

latentz with latentxz;

! regression equation;

y on latentx latentz latentxz;

MODEL CONSTRAINT:

! use labels from the MODEL section to impose constraints;

lxmean = 0;          

lzmean = 0;

lxzmean = lxmean*lzmean + covlxlz;

latxbyxz = zmean;

latzbyxz = xmean;

OUTPUT:

sampstat standardized;

Questions or suggestions? Email Craig Enders