Plots when modeling categorical data PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
Message/Author
 Jen posted on Tuesday, June 28, 2011 - 11:30 am
Hello,

I am modeling over 8 time points (equally spaced) with a dichotomous variable. A linear model is a very good fit according to all indices. However, when I look at plots of the actual and estimated probabilities, the estimated probabilities do not show a linear trajectory--instead, the trajectory bends at many points and follows the data quite closely.

The model syntax:
ix sx | copx2@0 copx3@1 copx4@2 copx5@3
copx6@4 copx7@5 copx8@6 copx9@7;

Is there any reason to expect this in the plots of models using categorical data?

Also, is there any way to know what intercept to use if I wanted to plot the estimated trajectory myself (since the intercept is 0 with categorical variables)?

Thank you!
 Linda K. Muthen posted on Wednesday, June 29, 2011 - 10:23 am
The growth model is linear in the logits or probits not the probabilities. However, I would not expect it to jump around. You can't compute it yourself if numerical integration is required.
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: