Page 1 of 1

Test of difference in R-squared

Posted: Wed Apr 04, 2012 10:44 am
by oksanakim
Hello guys,

I am estimating two regressions, for which independent and dependent variables are the same, except for the sample compositions and the time periods are different. Is there any way I could perform test of difference in adjusted R-squared for such regressions? Either manually or in Eviews? My understanding is that the F-test wouldn't work because I do not deal with "restricted versus unrestricted" regressions case. The seemingly unrelated regressions - I would also doubt that this is what I have here..... Any advice will be appreciated. Thank you.

Re: Test of difference in R-squared

Posted: Mon Apr 30, 2012 3:04 pm
by oksanakim
Hello again

I realized i might need to perform the Cramer's test, since I estimate the same model but for two different samples. Does anyone know how to model the Cramer's stats in Eviews? Thanks a lot.