Hi,

Is it possible to use CFA outputs, latent variables, as explanatory (independent) variables in a multinomial logistic regression model in a GSEM model? I thought this would be as easy as adding a path from the latent variable to my categorical dependent variable but it does not seem to work.

So what I do is I add a measurement component first, which works fine by itself, then I add a regression component, which works fine by itself as well, then add a path from the latent variable to the dependent variable in the regression component. But, I guess, rather than treating the latent variable as an explanatory variable for the regression, it treats my dependent variable as another source variable for the latent variable.

I hope this makes sense. I am normally an R user and an SEM newbie, so apologies if I am missing things that are obvious.

Ugur