Non-Normal Data and Bootstrapping PreviousNext
Mplus Discussion > Structural Equation Modeling >
 Alissa Banyasz posted on Tuesday, July 01, 2014 - 2:54 pm
Drs. Muhten,
I am using a data set with non-normal data and, therefore, have used MLR estimation to deal with this issue. I would also like to run bootstrapping to test indirect effects, but the BOOTSTRAP command cannot be run with MLR (only ML). (1)Is this because bootstrapping does not rely on the normality assumption and therefore does not "need" to be used in conjunction with MLR? (2)Also, is it appropriate to run my model with MLR to get parameter estimates and fit statistics, and then to rerun it with ML/BOOTSTAP in order to obtain the bootstrapping confidence intervals?
Thank you so much in advance!
 Bengt O. Muthen posted on Tuesday, July 01, 2014 - 5:17 pm
First note two things. Bootstrapping only affects the SEs, not the parameter estimates. MLR and ML only differ in how SEs are computed - the parameter estimates are the same.

(1) Bootstrap SEs is an alternative to MLR SEs.

(2) You can get MLR fit statistics, but parameter estimates will be the same.
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