|
|
Multiple group LCA with covariates |
|
Message/Author |
|
|
I have a LCA model with partial invariance. I would like to incorporate covariate(s) into the model using the 3-step approach, allowing the effects of the covariate(s) to vary across groups. Two questions: 1) The MPLUS output for a multiple group LCA model does not include the logits denoting classification accuracy needed to fix the logits in step 3. Why is this? 2) Using the 3 step approach with a multiple group LCA, what would the syntax be for fixing the logits within each group/class and allowing the regression of C on X to vary across groups? |
|
|
Treat the "multiple group" aspect as a second latent class variable with a perfect indicator and take a look at section 4 in http://statmodel.com/download/webnotes/webnote15.pdf which shows how this works with two class variables. Alternatively, you can follow the formulas in that note to manually obtain the logits, see formula (1) and section 3. |
|
Back to top |
|
|