Hello,
I am using the command margins after a logistic regression run with svy. The commands look like this:
Differently from the manual, the test I get does not present a chi2 distribution but an F one. If I ran the regression with no weights, I would have a chi2 distribution.
Could anyone explain why there is such a difference between weighted and unweighted tests?
Thank you and best regards,
Lydia
I am using the command margins after a logistic regression run with svy. The commands look like this:
svy, subpop(if sex==1): logit y x##z w
margins r.x, subpop(if sex==1) atmeans at(z=(1 2 3)) contrast(atcontrast(rb3))
margins r.x, subpop(if sex==1) atmeans at(z=(1 2 3)) contrast(atcontrast(rb3))
Could anyone explain why there is such a difference between weighted and unweighted tests?
Thank you and best regards,
Lydia
