Fixed effect and random effect model

[Akif Alig]

In Panel data analysis when should we use fixed effect model and when should we use random effect model?

[Carlo Lazzaro]

Akif:
too broad a question indeed.
You may benefit from reading any decent textbook on panel data econometrics (many are listed in the reference section of -xtreg- entry, Stata .pdf manual).

[Maxence Morlet]

If you're an economist, we tend to just throw two-way fixed effects at everything for causality purposes. See https://papers.ssrn.com/sol3/papers....act_id=3906345 for a comprehensive and excellent review.