Difference between displayed values when displaying invFtail and invchi2tail

[Jonas Rasmussen]

Hi as I understand it, an F-distribution should be asymptotically chi^2-distributed. However, when I run:
disp invFtail(3,3432,0.05)
I get roughly 2.61 while when I run:
disp invchi2tail(3,0.05)
I get roughly 7.81. Should Stata not display roughly the same result since 3432 degrees of freedom is sufficiently large?

[Ken Chui]

That seems right, the relationship is: chi = F * df. And 7.81 / 3 is about 2.60.

See https://www.stata.com/support/faqs/s...distributions/

[Jonas Rasmussen]

Thanks, yes I had forgotten that it scales with df.