|
|
MLM: non-normal & missing |
|
Message/Author |
|
Sung Kim posted on Monday, September 08, 2008 - 6:29 am
|
|
|
I am doing a path analysis and my data are not multivariate normal and have some missing (slightly less than 10%). I was trying to use MLM (Satorra-Bentler chi-square), but Mplus required to use LISTWISE=ON option. Why is that required? Is there any article to support and articulate the listwise deletion when MLM is used? Thanks. |
|
|
See the Yuan and Bentler paper. Missing data estimation has not been developed for MLM. You can use MLR instead if you don't want listwise deletion. |
|
Back to top |
|
|