I am having a problem with a model in which I am using the SUBPOPULATION command. The model is a CFA model with 2 factors and all of my indicators are binary. Type is complex, estimator is wlsmv. The errors I received are:
THE MODEL ESTIMATION TERMINATED NORMALLY ROBUST CHI-SQUARE IS NOT DEFINED DUE TO NONPOSITIVE TRACE.
THE STANDARD ERRORS OF THE MODEL PARAMETER ESTIMATES COULD NOT BE COMPUTED. THE MODEL MAY NOT BE IDENTIFIED. CHECK YOUR MODEL.
I have no reason to clearly attribute these errors to the subpopulation command, but I noticed that when I use this command my binary variables are actually being treated as 3 category ordinal variables. In the summary of categorical data proportions each item has a 3rd category with "0.000" proportion of cases. The associated threshold with this 3rd class is estimated as "5.791". Please let me know if this is supposed to happen with the subpopulation command, and why.
Also please note that the subpopulation command has appeared to have called out the correct subsample; that is, the number of observations is correctly listed as "1313".