Testing invariance with model constraint PreviousNext
Mplus Discussion > Categorical Data Modeling >
 X. Portilla posted on Friday, February 06, 2015 - 4:48 pm
I am interested in testing invariance in my model between two subgroups in my data. Ultimately, I am interested in understanding whether there are differences between the two groups, and whether the model is appropriate to use with the full sample.

I am considering two options:

1) Run a fully constrained model and then compare it to a fully unconstrained model, using Santorra-Bentler chi-square difference testing since I use MLR. If the difference is significant, follow up with MODEL CONSTRAINT to examine any individual pathway differences.

2) After running a fully constrained model, explore invariance of individual pathways using MODEL CONSTRAINT, irrespective of an overall difference in fit between a constrained and an unconstrained model.

Which option is preferred? Or is there a better solution?

Many thanks,
 Bengt O. Muthen posted on Saturday, February 07, 2015 - 2:48 pm
If you believe that most parameters are invariant I would start with a fully constrained model and then use Modindices to check for constraints that need to be freed up.

If you believe that few parameters are invariant I would start with the fully unconstrained model and add constraints to check which don't fit.
 X. Portilla posted on Monday, February 09, 2015 - 12:22 pm
Thank you for your prompt reply, Dr. Muthen!
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