Hello!

I could use some help with GEE models controlling for clustering. The schools were randomly assigned to intervention or business as usual and we have repeated measures (pre and post) for the outcome variable. We want main effects for time, treatment group, and interaction between time * treatment where the students are clustered/nested within school. Here is an outline of the syntax I'm using.

GENLIN outcome BY treatment time

/MODEL treatment time treatment*time

/REPEATED SUBJECT = student*school WITHINSUBJECT = time

I'm concerned that I am not controlling for the clustering correctly as I get the EXACT same results as when I use the following:

GENLIN outcome BY treatment time

/MODEL treatment time treatment*time

/REPEATED SUBJECT = student WITHINSUBJECT = time

Thanks for any thoughts or links on how to control for clustering/nesting in GEE.