Re: [Gretl-users] R-squared and the constant term (OLS)

Wednesday, 23 September 2009

2009/9/23 Allin Cottrell <cottrell(a)wfu.edu&gt;

...
 In the absence of an intercept gretl nowadays uses the formula
 recommended by NIST,

 R^2 = 1 - SSR/(sum of the y^2)

 Another possibility, which gretl used to report, is the square of
 the correlation between y and yhat. 

Allin, now everything is clear for me. EViews uses the second possibility
you mentioned (the square of
the correlation between y and yhat).

But I have another doubt: What about the adjusted R^2? How Gretl computes it
(in the "no intercept" case)?

Best,
Henrique