Quote:
Originally Posted by AbedNadir
the denominator of FTest appears to include the intercept, k is the parameter number and the denominator is nk, but the DoF in the model is k1

Good point, that’s why I listed out both of those two instances in one of my prior posts and I tried to equate them, but can’t cuz DoF just changes depend on context I think.
So DoF for a GLM is all beta exclude intercept aka beta0.
In F test, DoF is #datapoint minus #parameter(ALL bets include beta0) in the comparison step.
If anyone has any more updates on this please do so. Thanks!