Using time-varying predictor of laten... PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
Message/Author
 Anonymous posted on Monday, July 08, 2002 - 2:11 pm
I am running a growth mixture model using a y variable at 4 time points. I have been able to run this and to use time invariant variables to predict membership into different classes. However, I have not been successful with using a time-varying predictor. Some of my time varying predictors are across all 4 time points, while others are across 3 of the time points, and one is across only 2 of the time points. Any suggestions?
 bmuthen posted on Monday, July 15, 2002 - 2:58 pm
This type of modeling should work fine. It is hard to say anything more specific without analyzing your model. If you are a licensed user you may want to send your input, output and data to support@statmodel.com.
 Anonymous posted on Monday, April 11, 2005 - 9:46 pm
I am interested in a latent growth mixture model with 4 time points. A covariate would be time-variant. I have two questions: 1. Where can I find a specific example or paper of LCGM with time-variant covariates? 2. I can also model this covariate as several latent growth trajectories and relate them to the trajectories of y variable. How can I know which model is better? Can I still use AIC or BIC to compare the two models? Thank you!
 BMuthen posted on Wednesday, April 13, 2005 - 11:55 pm
1. I can't think of a paper to refer you to but the principles of single=-class analysis apply.

2. Yes, you can use BIC for this.
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: