Scale of factor loading and threshold PreviousNext
Mplus Discussion > Categorical Data Modeling >
 ffan posted on Wednesday, September 17, 2014 - 2:17 pm

I did two-group CFA with binary items using WLSMV estimator. Now I am looking at the factor loading and threshold. I read in a book chapter that these two can be analogous to IRT item discrimination and item difficulty, but they are not equivalent. As IRT can take both normal and logistic scales, I wonder what is the scale of factor loading and threshold in Mplus, respectively. Should I use the unstandardized solution or standardized when I report their respective values? And what is the right way to interpret both the unstandardized and standardized factor loading and threshold? Can you point me to an application paper or resources?

Thanks, Angie
 Bengt O. Muthen posted on Wednesday, September 17, 2014 - 3:45 pm
We have lots of documentation on our IRT page which you find through our home page's left margin:

WLSMV and Bayes estimation use probit while ML estimation can use probit or logit.

I would use standardized measurement parameters if you are just looking at a single group and time point.
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