www.cmm.bristol.ac.uk/forum

Hi,

I wonder if there exists some kind of thumb-rule for finding which level-1 predictors should be allowed to have heterogeneity in the random part. In the book "Developing multilevel models for analyzing contextuality, heterogeneity and change using MLwiN3" on page 78-79 it only says that increasing of standard error show that something is misspecified.
Thank you for the help.

Hi Rina,
Presumably you can do a likelihood ratio test between the models with and without heterogeneity and keep those predictors that make a significant difference in the random part?
Best wishes,
Bill.

I wonder if there exists some kind of thumb-rule for finding which level-1 predictors should be allowed to have heterogeneity in the random part