Time-varying covariates in discrete-t... PreviousNext
Mplus Discussion > Categorical Data Modeling >
Message/Author
 Michael Lemay posted on Thursday, January 08, 2009 - 1:33 pm
I am trying to run a discrete-time hazard model (Singer and Willett, 2003).
Currently, this is a very simple model I can fit with the logit command included
in most stat package. However, I plan to use Mplus to run more sophisticated
models. Before doing so, I want to be able to benchmark my Mplus results against the results I get with a more familiar package.

Here is my syntax:
Analysis:
type = mixture missing; algorithm=integration;
Model:
%overall%
evrout2-evrout34 on lseo (1);
evrout2-evrout34 on lmen (2);

This works just fine. However, problems arise when I try to include time-varying
covariates. (Both lseo and lmen are time-invariant) I thought that I could do
this by simply adding this line for the time-varying covariate: evrout2-evrout34
on linst2-linst34 (3); to the model statement. However, when I do so, I get an error message stating that I need to use the integration=Montecarlo option.
This makes me think that I'm doing something wrong.

Thank you for your help.

Michael
 Linda K. Muthen posted on Thursday, January 08, 2009 - 1:49 pm
Certain situations require Monte Carlo integration. Add INTEGRATION=MONTECARLO; to the ANALYSIS command. If you continue to have problems, please send your input, data, output, and license number to support@statmodel.com.
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: