What steps should be performed for Panel Estimation

[vtodorov]

Hello,

I would like to ask if anyone can guide me to a source or a book where I can read what are the steps I need to perform for a panel estimation.
Basically, I know how to go around the software (Version 6) and have set up my file for panel data (25 countries and 3 time periods).
But I have no idea whether or not should I run estimation without nay effects, with fixed effect, with random effect, with mixed effects and whether I should check any assumptions first (autocorrelation, heteroskedasticity).

So far I have run test equations with my data (no effects and cross-section random effects) and the results are both significant but only the no effects equation has a large F-statistic and strong r squared.

Thanks for any advice and help.
Vihren