Natasha,
Without answering your question, I seem to recall that consistent estimates of the parameter vector do not exist for fixed-effects probit models, but they do for fixed-effects logit models. This applies in general, so maybe the fixed-effects probit is not the best alternative. Perhaps a conditional logit model is more preferable (if you can live with the assumptions)?
Can anyone verify if my memory is correct?
Jan
-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of natasha agarwal
Sent: 29. juli 2011 10:48
To: statalist@hsphsun2.harvard.edu
Subject: st: Fixed Effects Probit Model.
Dear All,
I have a firm-level unbalanced panel data set (20,000 firms for 5
years), and wanted to run a two-way fixed effect probit model.
