Bracketed Factor? PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
 Jim Hamilton posted on Friday, August 02, 2013 - 12:27 pm
Hi there,

A colleague generously donated some syntax she used to test trait stability. The model is below. My specific question is, what is the effect of specifying the time two latent measure of the trait as:




con1 by
C11* (1)
C12 (2)
C13 (3);

con1 @ 1;

con2 by
C21* (1)
C22 (2)
C23 (3);

[con2 ];

sexint |gendercent XWITH con1;
pharmint | pharcent XWITH con1;
threeway |sexint XWITH pharcent;

con2 ON con1 gendercent pharcent sexint
pharmint sexpharm threeway;

C11 C21 (4);
C12 C22 (5);
C13 C23 (6);

C11 WITH C21 ;
C12 WITH C22 ;
C13 WITH C23 ;

[C11 C21](7);
[C12 C22](8);
[C13 C23](9);
 Bengt O. Muthen posted on Friday, August 02, 2013 - 1:27 pm
That frees the factor mean at time 2. By default single-group models have factor means fixed at zero. But in a longitudinal model with measurement invariance such as yours, the factor mean differences across time are identifiable. So time 1 factor mean is zero and time 2 factor mean is free and can be compared to the zero of time 1.
 Jim Hamilton posted on Friday, August 02, 2013 - 1:38 pm
Great. Thanks.

And may I also say thanks, in general, for your patience and helpfulness with all our questions, sophisticated and not-so-sophisticated.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message