Latent Growth Curve Model with Modera... PreviousNext
Mplus Discussion > Growth Modeling of Longitudinal Data >
 Andrew Triplett posted on Thursday, February 18, 2016 - 11:23 pm

I am working on a project and need help with how to set up the syntax. We are looking at longitudinal data, specifically the interaction of two growth curves on the growth curve of an outcome. For instance, how the trajectory of life satisfaction moderates the relationship between the trajectory of financial status and the trajectory of happiness. We are expecting that if life satisfaction is improving over time, then the even if financial status is worsening, the improvement of life satisfaction will mitigate the effects on happiness. How would I go about setting up a model to test these types of interactions?
 Bengt O. Muthen posted on Friday, February 19, 2016 - 6:19 am
Sounds like you want interactions between growth factors so use XWITH. The UG index will show you examples and there are also FAQs on "latent variable interactions".
 Andrew Triplett posted on Friday, February 19, 2016 - 6:49 am
Thank you so much for your help! Could you by any chance post the link to the relevant example for my model?
 Bengt O. Muthen posted on Friday, February 19, 2016 - 6:22 pm
FAQs are at

and the ones you want are called

Latent variable interactions
Latent variable interaction LOOP plot
 Andrew Triplett posted on Wednesday, March 02, 2016 - 7:52 pm
Hi Dr. Muthen,

I think I have figured out the syntax and it appeared to run right, but I am not getting any indices of model fit, does this syntax look correct or am I missing something?

names = ID PD1 PD2 PD3 PD4 LS1 LS2 LS3 LS4 FS1 FS2 FS3 FS4 Happy1
Happy2 Happy3 Happy4 SE1 SE2 SE3 SE4 Health1 Health2 Health3
Health4 Inc_Needs2 Inc_Needs3 Inc_Needs4 Dis1 Dis2 Dis3 Dis4;

usevariables = LS1 LS2 LS3 LS4 FS1 FS2 FS3 FS4 Happy1 Happy2
Happy3 Happy4;
Missing = all (999);

Analysis: TYPE = RANDOM;

f1 BY LS1-LS4;
f2 BY FS1-FS4;
f3 BY Happy1-Happy4;
f3 ON f1 (b1);
f2 (b2);
f1xf2 | f1 XWITH f2;
f3 ON f1xf2 (b3);

sampstat residual standardized TECH1 TECH8;
 Linda K. Muthen posted on Thursday, March 03, 2016 - 7:00 pm
Chi-square and related fit statistics are not available when means, variances, and covariances are not sufficient statistics for model estimation. This is the case with numerical integration. Nested models can be tested using -2 times the loglikelihood difference which is distributed as chi-square.
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