Use of LC probabilities to predict ou... PreviousNext
Mplus Discussion > Latent Variable Mixture Modeling >
Message/Author
 Jason Downer posted on Tuesday, February 15, 2011 - 1:53 pm
We have conducted a LCA and have decided on a 3-class solution. We have used the savedata option to save the posterior class probabilities. How exactly do you go about using the class probabilities as predictors of a distal outcome, and how would you interpret those results (i.e., the coefficients associated with the class probability variables)?
 Bengt O. Muthen posted on Tuesday, February 15, 2011 - 2:08 pm
You can do the analysis with the distal in a single step together with the LCA.

If you want to do it in a single step, you could either use most likely class as a predictor or use the class probability as a continuous predictor. For this to work well, your entropy should be high. See also the dissertation on our web site:

Clark, S.L. (2010). Mixture modeling with behavioral data. Doctoral dissertation, University of California, Los Angeles.
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: