Zero-class in LCGA with categorical o... PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
Message/Author
 Jacqueline Homel posted on Thursday, December 06, 2012 - 3:31 pm
I am running a latent class growth analysis of frequency of cannabis use from age 14 to 25, where the cannabis use variable is ordered categorical (five categories). I have specified an a priori zero-class like this:

%c#2%
! zero class ;
i@0 ;
s@0;
q@0 ;
[i-q@0] ;

However, the thresholds in this class are still the same as the other classes. Should I also set the thresholds to zero, and if so, how is this done? Thanks a lot.
 Bengt O. Muthen posted on Thursday, December 06, 2012 - 4:10 pm
For your 5-category outcome u, in class 2 you can say

[u1$1@15 u1$2@16 u1$3@17 u1$4@18];

so that the probabilities of exceeding these high thresholds are zero while the thresholds are still ordered. Then do that for u2, u3, etc for the outcomes at all time points.
 Jacqueline Homel posted on Tuesday, December 11, 2012 - 4:02 pm
Thank you very much! This helped a great deal.
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: