Chi-square difference test is overpow... PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
Message/Author
 Jane W. posted on Tuesday, January 08, 2013 - 5:41 am
I'm testing a 4-factor solution against a second-order CFA with 4 lower-order factors. My sample sizes range from 1,000-9,000+, so I'd imagine that the chi-square diff tests are overpowered and much too sensitive to trivial differences.

My questions:
- Are there any large-sample alternatives to the chi-square difference test?
- Or can I simply adopt a more stringent criterion (maybe p<.001 or p<.0001)?
- Finally, can you point me to any articles on this topic?

I appreciate any suggestions! Thank you!
 Bengt O. Muthen posted on Tuesday, January 08, 2013 - 5:58 pm
- Not really. You can try working with BIC

- Becomes an arbitrary choice

- One way to go is to work with less restrictive models as in

Muthén, B. & Asparouhov, T. (2012). Bayesian SEM: A more flexible representation of substantive theory. Psychological Methods, 17, 313-335.

which is on our web site.
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: