Re: st: areg vs xtreg, fe

["jenny.montaldo" <jenny.montaldo@libero.it>]

Thank you Dave

I also think that the R^2 should not matter so much, since it is not what I am interested, but referees sometimes do it. Anyway, I will report the R^2 within as you suggested me, cause I also know that stata in that case maximize the R^2 within. 

Do you know something about the second question? That if i run xtreg, fe for white collars there is no problem, while with areg it says that the matrix is not positive definite??

Many thanks
jenny

---------- Initial Header -----------