Bootstrapping option for Type = twole... PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
Message/Author
 Charlotte Vrijen posted on Friday, February 17, 2017 - 3:33 am
Because I suspect non-normality I would like to repeat my multilevel analyses by using a bootstrap method. Is it possible to do bootstrapping when using Type = twolevel random and Estimator = MLR?

In case this is not possible, I also read that MLR already produces p-values that are robust to normality violations. Is this correct and is it described in a paper?

Many thanks!
 Bengt O. Muthen posted on Friday, February 17, 2017 - 10:01 am
I think you are referring to nonnormality of your variables as opposed to nonnormality of your effects (such as an indirect effect a*b). MLR is suitable for the former but not the latter.

Regarding MLR SEs, look for articles by Satorra & Bentler and ask on SEMNET.
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: