MLM or MLR estimator for CFA PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
 seungbin oh posted on Tuesday, April 24, 2018 - 7:37 am
Dear Drs. Muthén and Muthén,

I am conducting CFA to verify the factor structure and psychometric features of my own instrument. And, I have a quick question about estimators, specifically MLM and MLR. My instrument consists of 23 items and uses a 5 Likert-scale. No missing data exists, but the assumption of multivariate normality is violated. Thus, I've been trying to figure out whether MLM or MLR should be used for my CFA. In my limited knowledge, both estimators could be used for my CFA and robust to the non-normality of data.

I tried both, and the results were similar with MLM estimator producing a little bit better CFI values and chi-squire values.

Any insight would be most appreciated!
 Bengt O. Muthen posted on Tuesday, April 24, 2018 - 2:30 pm
MLR is my favorite. You may want to ask Alberto Maydeu to send you (us) his latest version of a paper where he makes these comparisons:

Maximum likelihood estimation of structural equation models for continuous data:
Standard errors and goodness of fit
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