Modeling interaction effects in LGCA PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
Message/Author
 Nina Wu posted on Monday, October 17, 2011 - 8:20 pm
Hi. I am writing to ask if interactions can be modeled in LGCA. For instance, if I want to interact maternal education and mothers' race/ethnicity to predict high chronic child outcome, would I be able to model that in LCGA?

Any help would be appreciated. Thank you in advance.

Nina
 Linda K. Muthen posted on Monday, October 17, 2011 - 10:21 pm
Yes, you can create an interaction term and use it as a predictor of the latent classes.
 Nina Wu posted on Thursday, October 27, 2011 - 5:15 pm
Thank you for your reply, Dr. Muthen. I would like to ask a follow-up question.

Let's say that I have identified 3 different latent classes of a variable: high/low/moderate, can I take a step further and test for differences by race/ethnicity and immigration status?

Also, I was told by a colleague that I would be able to model co-occurrence (2 behaviors together) in LGCA. Is that true?
 Bengt O. Muthen posted on Friday, October 28, 2011 - 12:26 am
Q1. Yes, you include those variables as predictors of the latent class variable.

Q2. Yes, that is corect. Your 2 processes can be influenced either by one and the same latent class variable, or have their own latent class variable, letting the 2 latent class variables correlate.
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: