SEM with clustered data PreviousNext
Mplus Discussion > Structural Equation Modeling >
 anne C posted on Thursday, September 27, 2012 - 12:22 pm
Dear all,

I am attempting to run a model using Mplus, to which I am not very familiar.
The idea is to compute direct and indirect effect of 4 outcomes (binary) on various predictors using a dataset which has family structure (individual-level are child, nested into 1 parent).

I have done this

TITLE: SEM attempt with clustered data;
DATA: FILE =datos.dat;
VARIABLE: NAMES = iid centre num ac1 ac2 ac3 ac4 ap
sex agp1 agp2 ;
usevariables = iid ac1 ac2 ap agp1 ;
missing are all (-9999);
CATEGORICAL = ac1 ac2 ap ;
CLUSTER is iid;


ac1 ON ap agp1 ;
ac2 ON ap agp1 ;
ap ON agp1 ;

ac1 IND agp1;
ac2 IND agp1;

SAVEDATA: estimates=e:\sem1_estim.dat;

The program runs well but I would like to be sure that I am doing the right thing here.
Does the structure of the data is well-taken into account with the CLUSTER option? How are the standard-errors computed? Is there a better way to do this?
Thank you very much in advance

 Linda K. Muthen posted on Thursday, September 27, 2012 - 8:49 pm
Do the child and parent have the same iid?
 anne C posted on Friday, September 28, 2012 - 8:05 am
Dear Linda,

The iid identifies the parent, while the child is identified, within the same parent, by the order of birth (num=1 for first child, 2 for second etc).
I used the "cluster iid" to take into account the familial structure since in the model, ag1 and ap are variables at parental-level (common for children of the same parent).
Thanks for your help


 Linda K. Muthen posted on Friday, September 28, 2012 - 3:51 pm
Then your analysis looks like it is set up correctly.
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