0
$\begingroup$

I have a question regarding the Relationship between R squared and F-Test under heteroskedasicity.

If R squared is uaffected by the heteroskedasicity, since it is a population estimate, then why is F-stats invaild, because F-test can be conducted using R squared.

Thanks

$\endgroup$

1 Answer 1

0
$\begingroup$

I think I understand now, after reading Textbook. R squared = 1 - var(u)/var(X), however, under heteroskedastic error, these two estimate are not vaild for finite sample distribution, as the estimated error variance are not vaild in this case.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.