Group test of equality for variance
Posted: Sat Jan 12, 2019 6:59 pm
Hi,
I was running an F-test for the equality of the variances of two populations based on two small independent samples. I used EViews 8.1 and the relevant workfile is attached.
My understanding is that EViews calculates the F-ratio as the square of s(L)/s(S), so the corresponding degrees of freedom should be n(L)-1 for the numerator and n(S)-1 for the denominator. In my example investment_2 has larger standard deviation than investment_1 and the test statistic is (23.02638/9.06826)^2=6.44768. The corresponding degrees of freedoms are 7 (numerator) and 9 (denominator), but EViews reports (9,7). Why?
Thanks for your help.
Laszlo
I was running an F-test for the equality of the variances of two populations based on two small independent samples. I used EViews 8.1 and the relevant workfile is attached.
My understanding is that EViews calculates the F-ratio as the square of s(L)/s(S), so the corresponding degrees of freedom should be n(L)-1 for the numerator and n(S)-1 for the denominator. In my example investment_2 has larger standard deviation than investment_1 and the test statistic is (23.02638/9.06826)^2=6.44768. The corresponding degrees of freedoms are 7 (numerator) and 9 (denominator), but EViews reports (9,7). Why?
Thanks for your help.
Laszlo