WLS vs ULS for ordinal items? PreviousNext
Mplus Discussion > Exploratory Factor Analysis >
 Bruce A. Cooper posted on Thursday, January 05, 2012 - 9:27 am
I have lots of gaps in my basic knowledge, and this is one. Hope you can help!

In doing EFA (and other analyses) with binary or ordinal items, I want to compare models using ULS and WLS (and their variants) to decide which estimator works best. Present sample size is about 380 with many skewed 0-4 ordinal items, and I may be using between 80 and 100 items.

I'd appreciate any references you can offer that would help fill in this gap about how to choose the best estimator, vis a vis measurement scale, sample size, generalizability, and model complexity!

Thanks, and Happy New Year!
 Linda K. Muthen posted on Thursday, January 05, 2012 - 12:21 pm
This type of decision can be guided by using a Monte Carlo study. Albert Maydeu has done papers in the SEM journal on this topic and Flora and Curran have done an article in Psych Methods.
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