Multivariate nonnormality PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
Message/Author
 Ioanna Vrouva posted on Saturday, November 07, 2009 - 9:06 am
I would be very grateful for your guidance with the following. I have run a CFA with categorical variables (WLSMV estimation).
Data were nonnormal at univariate level.
I have been asked by a Referee whether any multivariate nonnormality indices were examined. Does Mplus offer this and if so how can one obtain these indices?

I was under the impression that when there is univariate nonnormality, it can be assumed that the data are nonnormal also at a multivariate level. I would also write in my reply that : ...
"The Weighted Last Squares Approach implemented in MPlus is robust with sample sizes 200 or larger (Muthén, du Toic & Spisic, 1997), and MPLus applies scaling corrections (uses scaled chi-square) in the fit indices calculation".

Is this correct?
Thank you very much
Ioanna Vrouva
 Linda K. Muthen posted on Saturday, November 07, 2009 - 11:56 am
For categorical outcomes, the issues of non-normality and skewness are not relevant. Categorical data methodology is developed to correctly analyze categorical variables with floor or ceiling effects.
 Ioanna Vrouva posted on Saturday, November 14, 2009 - 7:06 am
Thanks very much
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: