For general information, see http://www.stata.com/manuals14/xtxtreg.pdf p. 11. For formulae, see the same document, pp. 27-29. See also the discussion of examples there.

In your case, the directly relevant goodness of fit measure is the R-sq. between (0.8188), unless you want to predict the within model or fit the overall data.

For xtreg, be, and for xtreg, fe, Stata saves the value of adjusted R-squared in e(r2_a), so that after running xtreg, you can simply write dis e(r2_a) do display its value. But see http://www.statalist.org/forums/forum/general-stata-discussion/general/1290416-anyone-knows-what-is-an-adjusted-within-r2 on the discussion whether adjusted R-squared is useful in this case at all.

Great explanation, thank a lot 

Dear Annarita Macc,

You can use this code that I have used in my research:

xtreg dep_vble indep_vble_01 indep_vble_02 ... indep_vble_n temporal_vble, fe

estimates store fix

xtreg dep_vble indep_vble_01 indep_vble_02 ... indep_vble_n temporal_vble, re

estimates store ran

hausman fix ran, sigmamore

* If P-value < 0.05 -> H0 is rejected, FE are used because they are consistent and efficient

* If P-value > 0.05 -> Can not reject H0, the RE are used because they are consistent

Best Answer:

My dear, attached file will help you a lot...