TITLE: ESTIMATING CHANGE IN CRITERION VALIDITY
DATA: FILE = ;
VARIABLE: NAMES = UB UA C; ! SCALE SCORE BEFORE AFTER REVISION, CRITERION.
MODEL: F1 BY UB*1;
F2 BY UA*1;
F3 BY C*1;
UA-C@0;
F1-F3@1; ! --> LATENT CORRELATIONS = LATENT COVARIANCES!
F1 WITH F3 (P1);
F2 WITH F3 (P2);
MODEL CONSTRAINT:
NEW (DELTA); = LOSS OR GAIN IN CRITERION VALIDITY DUE TO REVISION.
DELTA = P2-P1;
OUTPUT: CINTERVAL; ! REQUESTS CI OF DELTA.
Note. With deviations from normality (not resulting from ceiling/floor effects, clustering effects, or highly discrete components), add the line ANALYSIS: ESTIMATOR = MLR; to invoke the robust ML method of estimation. Check output section on confidence intervals, for that of the change D in criterion validity, at confidence levels 95% and 99%. (If a different confidence level is to be used, compute the CI for D by adding and subtracting its large-sample standard error produced by this Mplus command file, multiplied by the pertinent quantile of the standard normal distribution.)