Estimator for logit reg with missing ... PreviousNext
Mplus Discussion > Categorical Data Modeling >
Message/Author
 Qi posted on Friday, July 01, 2011 - 10:46 am
Dear Dr. Muthen,

I am conducting a logistic regression with missing data on predictor. In example 3.5, the ESTIMATOR=ML specificaiton is given to run a logistic regression. I specified the variance of the predictor in the model command so Mplus can impute data for IV. I noticed that "WLSMV" was the default estimator and I was able to get results for the whole sample. However when I specified other estimators such as ML or MLR, I got the following message:

*** FATAL ERROR
THIS MODEL CAN BE DONE ONLY WITH MONTECARLO INTEGRATION.

Does it mean I can only use WLSMV when imputing for missing IV? Is the WLSMV result trustable logistic regression output?

Thanks!
Qi
 Qi posted on Friday, July 01, 2011 - 11:00 am
btw, my DV has no missing data and is categorical, and the IV has missing data and continuous/categorical
 Linda K. Muthen posted on Friday, July 01, 2011 - 2:24 pm
This means you need to add INTEGRATION=MONTECARLO; to the ANALYSIS command when you use maximum likelihood.

WLSMV is probit regression not logistic regression.
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: