Maximum likelihood scale score analysis

DATA:

file = eatingrisk.dat; 

VARIABLE:

names = abuse bmi bds1 - bds7 edr1 - edr6;

usevariables = abuse bds2 - bds7 edr1 - edr5 bodydis eatrisk;  ! select variables;

missing = all(-99);

DEFINE:

bodydis = sum(bds1-bds7);  ! compute scales;

eatrisk = sum(edr1-edr6);

ANALYSIS:

estimator = ml;

MODEL:

eatrisk on abuse bodydis;

abuse bodydis;  ! include incomplete predictors;

! k-1 scale items as auxiliary variables;

bds2 - bds7 edr1 - edr5 with abuse bodydis eatrisk;  ! auxiliary variable correlations;

bds2 - bds7 edr1 - edr5 with bds2 - bds7 edr1 - edr5;

OUTPUT:

standardized(stdyx) sampstat;

Questions or suggestions? Email Craig Enders