|
|
Examinng model fit with latent modera... |
|
Message/Author |
|
|
Hello! I am testing several competing SEM models, where the baseline model does not have moderators, and the alternative models have one or two paths moderated by a latent construct. The moderated models employ the xwith syntax, and I noticed that much of the output I typically interpret is not available (e.g., indices to compare model fit). Where might I read up on an output option or syntax to incorporate to compare the the fit of models requiring 'xwith'? Thank you for your assistance! |
|
|
Model fit indices for models with XWITH have not yet been developed in the literature. I would make sure your model fits well without XWITH and then simply see if the slopes for XWITH variables are significant or not. You can also use BIC. |
|
Back to top |
|
|