If you use stata

You have to use the following commands

xtreg y x1 x2, fe

With y : the dependent variable

And x1 x2 : the independent variables.

If p value

Thank you for your answer. If I find that at the begining p value > 0 .05, so the OLS is more appropriate than fixed effects model. So what should I do after ? Run an OLS analysis ? knowing that my independant variable is a dummy variable. Thank you again

Yes in that case OLS is more appropriate.

Yes for sure

Onipe suggestions is in line because you have to inform STATA that you are using panel data. OLS should be the first analysis. To have a robust and reliable analysis you then go further to analyse the data on FE and RE. You will now compare the result of the two to know which one is better.

Thank you very much. 5 years is not short to use Panel data ?

Provided u have annual observations of 5 years you may use fixed and random effects by comparing the results, as was pre told.

However, what you can NOT check for robustness since Generealized Method Of Moments does not apply dut to the small number of obs.

So I can use panel data if I use fixed or random effects even if I have the small T (5years)?

Yes you may.. but without being able to check for robustness with GMM.

is it obligatory to check for robustness with GMM ? Thank you very much

is it obligatory to check for robustness with GMM ? Thank you very much