Re: Fixed effects or random effects: The Mundlak approach

Hi. I was wondering if you have a citation for the assertion "Unlike the latter, the Mundlak approach may be used when the errors are heteroskedastic or have intragroup correlation." Thanks you in advanced