Moderators in GMM models - is it poss... PreviousNext
Mplus Discussion > Latent Variable Mixture Modeling >
Message/Author
 Irina Patwardhan posted on Wednesday, November 15, 2017 - 7:39 am
I am estimating a longitudinal GMM, where i am interested first, in a number of trajectories in children's executive function (based on their rate of change), and second, how children's rate of change in executive function in elementary school is associated with different achievement outcomes in middle school (e.g., distal outcomes).

I am wondering if it is possible to include continuous moderators in these analyses? something like creating an interaction variable, moderator X executive function, or should it be moderator X class?

I have only seen discussions about GMM models with the distal outcomes and covariates, but nothing about moderators. Should i treat my moderators as covariates then?
 Bengt O. Muthen posted on Wednesday, November 15, 2017 - 1:14 pm
That should be based on substantive considerations. Perhaps you want to allow

moderator*s

where s is the executive function trajectory slope. This would use XWITH.
 Irina Patwardhan posted on Thursday, November 16, 2017 - 8:39 am
Thank you, Bengt! Can you guide me to any publications where people have done something like that?

Irina
 Bengt O. Muthen posted on Thursday, November 16, 2017 - 3:38 pm
Just say in your model command (using made-up names):

int | mod XWITH s;

y on s mod int;
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: