Extract R-square in panel data regressions

[Rochelle Zhang]

Dear all,

I have used the following statement to save R-square in panel data regressions, some of the R-square. My understanding is that R-square can't be negative.

Could someone point out if my thinking is wrong or the code I use was wrong.

egen both = group(firm year)

statsby _b _se r2 = e(r2_a), by (both) saving(Rsquare): regress y x1 x2 x3 , robust

thank you,

Rochelle

[Clyde Schechter]

R squared can't be negative, but e(r2_a) is adjusted R squared, which can be negative. If you want R squared, use e(r2) instead.

[Rochelle Zhang]

Dear Clyde,

You are always so helpful !!! Thank you!!!

Rochelle