LCA: covariates and multuple groups PreviousNext
Mplus Discussion > Latent Variable Mixture Modeling >
Message/Author
 Michelle Martel posted on Wednesday, February 25, 2009 - 1:09 pm
A reviewer has asked me to show factorial invariance for sex and age in a latent class analysis using continuous predictors. As a preliminary step, I was attempting to assess whether sex or age was a significant covariate in the models.

My understanding as to how to do this is to regress the latent class on sex/age (e.g., c ON sex). However, I keep getting error messages that c is an unknown variable. I'm puzzled as to why I am getting this error since it seems like this should work (manual example 7.12).

Also-- if I do successfully run this analysis and the covariates are significant, then is my best bet to run the multiple group analysis as per example 7.21 in the manual?

Thanks a lot!

Michelle
 Linda K. Muthen posted on Thursday, February 26, 2009 - 9:57 am
To test for invariance you would regress the factor indicators on x not the categorical latent variable on x. The statement c ON x probably does not work because you are using a version of the program in which this shorthand statement was not yet implemented.

You test for measurement invariance across classes in the same way you test across groups. See the Topic 1 course handout for the models to sue.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: