Estimated Probabilities as function o... PreviousNext
Mplus Discussion > Latent Variable Mixture Modeling >
 Frank Germann posted on Thursday, July 22, 2010 - 6:03 pm

My model (LCA w/ covariates) has 3 observed continuous dependent variables, 3 latent classes (as per MPLUS), and 6 covariates (of which 5 are categorical and 1 is continuous). I am currently using MPLUS Version 3.0.

Here are my questions:

1.) When I select the estimated prob. as function of covariates plot to look at the relationship between class probability and my covariates, it seems that MPLUS only considers the first three covariates entered when drafting the plot(although all of the covariates are listed under "values for x" in the window that pops up after selecting the plot type). Also, MPLUS will let me fix the (e.g.) 4th covariate listed under "values for x" to (e.g.) "1" but then nothing changes (for, I think, the reason mentioned above). Am I missing something? Would this be different in a newer version of MPLUS?

2.) I noticed that the class membership information provided in the plot (i.e., class 1, 27.7% etc.) doesn't change when I fix a covariate to (e.g.) "1". For example, if I had both males and females in my data but then wanted to see how (e.g.) the probability of males being in each one of the classes varies as a function of age (another covariate), shouldn't the percentage of class membership change since I'm now only looking at males?

Thank you!!
 Linda K. Muthen posted on Thursday, July 22, 2010 - 8:24 pm
I would need to see the relevant outputs and gph files along with your license number at to answer your questions.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message