Testing for group differences in SEM PreviousNext
Mplus Discussion > Structural Equation Modeling >
 Nina Wu posted on Tuesday, May 01, 2012 - 1:15 pm

I would like to know if I can test for group differences in SEM (as I can in growth curve modeling by constraining the intercepts and slopes across groups) by using constrains. If so, what are the parameters I would need to constrain?

Thank you,

 Linda K. Muthen posted on Wednesday, May 02, 2012 - 9:40 am
Yes, you can do this. You would constrain the parameters you are interested in testing. The multiple group section of the Topic 1 course handout shows examples of testing means, variances, and covariances of factor across groups.
 Nina Wu posted on Wednesday, May 02, 2012 - 12:35 pm
Thank you very much for your previous response, Dr. Muthen.

In growth curve modeling, I would constrain the intercepts and slopes across groups. May I ask what parameters I would need to constrain across groups in SEM?

a typical constrain statement in growth curve is:

model: [i] (1); [s] (2)

however, in SEM the model estimates the paths, not parameters such as intercept and slope. May I ask what is/are typical parameters that researchers constrain and how I would go about doing that?

 Linda K. Muthen posted on Thursday, May 03, 2012 - 11:37 am
In SEM structural parameters are means, variances, covariances, and regression coefficients. Any of these can be compared. I think regression coefficients are usually of the most interest.
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