Using vars with differing numbers of ... PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
Message/Author
 Jessica Schumacher posted on Monday, October 15, 2007 - 10:46 am
Hello,
I am running a One-Factor CFA using a series of ordered categorical variables. I specify that the variables are categorical in my code and am using the wlsmv estimator. My question is the following -- some of my variables have 6 ordered categories while others have 4. I have read in prior posts that the default correlation matrix computed in MPLUS with this type of data is the polychoric correlation. Is this correct? Also -- is it the case that the CFA is being performed on the correlation as opposed to the covariance matrix? If not, is there a way to specify this in MPLUS? Thank you for your help.
Jessica
 Linda K. Muthen posted on Monday, October 15, 2007 - 2:50 pm
With weighted least squares estimation and ordered categorical variables, the sample statistics used for model estimation are polychoric correlations.
 Jessica Schumacher posted on Monday, October 15, 2007 - 4:04 pm
I am assuming then that it is appropriate to use variables in the CFA that have differing numbers of ordered categories. Thank you very much for getting back to me so quickly!!

Jessica
 Linda K. Muthen posted on Monday, October 15, 2007 - 4:23 pm
Yes, it is.
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: