Comparing non-nested models PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
Message/Author
 Jacob Keech posted on Thursday, December 01, 2016 - 5:33 pm
I am using ML to estimate two path models that are identical, except that I am using Measure A of an IV in one model, and a different measure (Measure B) of the same IV in the second model. I want to decide which measure is superior and am wondering if it is appropriate to use conventional model fit statistics in this comparison; or, if there is an analysis where Mplus can compare both models?
 Linda K. Muthen posted on Friday, December 02, 2016 - 11:29 am
If the two models have the same set of dependent variables and the model is estimated conditioned on the independent variables, you can compare the models using BIC.
 Jacob Keech posted on Thursday, December 15, 2016 - 9:57 pm
Thanks Linda. In the model I also have two control variables with paths predicting the measure I am comparing (and all other variables).

Including or excluding the control variables changes which model has the best fit (according to smaller BIC) with this data. Would it be appropriate to exclude control variables for this type of model comparison?
 Linda K. Muthen posted on Friday, December 16, 2016 - 6:07 am
Please send the two outputs and your license number to support@statmodel.com.
 Angel Arias posted on Tuesday, January 09, 2018 - 5:06 pm
Hello Drs Muthén,
Is there a procedure in Mplus to do model difference testing of non-nested models with categorical indicators? Any advice?

Best,
Angel
 Bengt O. Muthen posted on Tuesday, January 09, 2018 - 5:31 pm
You can use ML and BIC.
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: