Hi Statisticians,
Is any one known the difference of GEE and regression with robust error variance?
For example, we consider binary outcome grouped data (r is the events, n is the total, group is intervention status, cat is stracified categories), modeled with GEE and GLM with robust error variance, as:
***GEE
stset cat
xtgee r group, family(binomial n) link(logit) eform

***GLM
glm r group, family(binomial n) vce(cluster cat) eform Thanks.