ULS for SEM PreviousNext
Mplus Discussion > Structural Equation Modeling >
 Maren Winkler posted on Monday, September 21, 2009 - 2:22 am

I'm estimating a model with 135 variables and 1187 subjects. Since the majority of variables is categorical, I've used the WLSMV estimator. During the calculations, Mplus suggests to use the ULS estimator instead. Now I've tried that but got the following message:
"Estimator ULS is not available for analysis with individual data.
Default estimator will be used."

What kind of data does the ULS estimator need?
why does Mplus suggest to use this estimator but then does not use it when I require it?

Thanks for your help!
 Linda K. Muthen posted on Monday, September 21, 2009 - 7:47 am
ULS is now available only for EFA. You can use ULSMV but I'm not sure if it will be helpful as it estimates both parameter estimates and standard errors.
 Maren Winkler posted on Tuesday, September 22, 2009 - 2:06 am
in which situations would ULSMV be more appropriate than WLSMV?
 Linda K. Muthen posted on Tuesday, September 22, 2009 - 10:18 am
ULSMV might be better with small samples.
 Bernd posted on Tuesday, January 19, 2010 - 12:50 am
Dear Linda K. Muthen,

I would be grateful if you could provide a reference showing that "ULSMV might be better with small samples."

Thanks a lot for your help.

 Linda K. Muthen posted on Tuesday, January 19, 2010 - 8:18 am
I don't know of a reference.
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