Excessice multivariate Kurtosis PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
 Alexander Munz posted on Saturday, June 13, 2009 - 6:52 am
Dear reader,

I am trying to fit a CFA-Modell to questionnaire date of translated version of a questionnaire, which includes 5 correlated Factors and some 38 Indicators (I know this are lots of indicators but thats not the issue here, since I do not want to parcel.)
My main problem is, that although there is only small univariate skewness and kurtosis in the data, there is excessive multivariate kurtosis (z=66.3). I don`t have the N for ADF or anything related (N=273, no missing Data, and couldnt find a way to procede. Should I use ML or GLS/OLS or use the storra & Bentler scaled approach what do you propose

Kind regards & thanks in advance Alexander
 Bengt O. Muthen posted on Sunday, June 14, 2009 - 10:33 am
Use MLM - it takes into account the non-normality in estimating standard errors and chi-square. MLM gives chi-square by Satorra-Bentler. If you had missing data, you should use MLR.
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