I've got a beginner's question about pooled Ols versus fixed effects. If I
run a fixed effects regression (and random effects have already been
excluded) and get an F test at the end of the output saying that u_i are
zero, does that mean that I should run a pooled OLS regression instead (or
do I have some more serious problem then)?
Moreover, how can I then test for autocorrelation in the pooled OLS model,
as none of the dwstat etc. tests work?
With regard to the problem of downloading ado-files: I tried to do that
yesterday and now I have to have Stata re-installed, as I can't run
regressions anymore. I'm not entirely sure whether that's due to the
downloading, but it might be a possibility reading the previous mails about
downloading problems.
