Duru Alan posted on Friday, September 28, 2012 - 12:02 pm
I am running a Multigroup CFA on categorical (ordinal) indicators as follows:
MODEL: socioeco by school workplac lookjob; public by neighb street; control by goingout police; [socioeco@0public@0control@0]; Model G2: [socioeco* public* control*];
When I donít add the group specific command and set factor means to 0 in both groups. I get the following error message:
THE LATENT VARIABLE COVARIANCE MATRIX (PSI) IN GROUP G2 IS NOT POSITIVE DEFINITE. THIS COULD INDICATE A NEGATIVE VARIANCE/ RESIDUAL VARIANCE FOR A LATENT VARIABLE, A CORRELATION GREATER OR EQUAL TO ONE BETWEEN TWO LATENT VARIABLESÖ
1) I wonder why I donít get the same warning, when I free factor means in G2, although I have one factor correlation over .95 in both models?
2) And what does this imply, my group means are different?
3) Would the model still be acceptable with this warning if I have theoretical justification and model fit is good?