|
|
Hi there.
For the purpose of a cross-country comparison, I am running separate
logistic regression for each country with the same variables each. My
question is, is it statistically legitimate to compare the size and
direction of the same coefficient across countries? I have read of a test
analogue to the Chow test for logistic regression (DeMaris 2004), but this
test assumes that error variance is the same across groups (i.e. countries,
in my case). According to Allison 1999, this is rather unrealistic to
assume, however I'm not quite sure how to test for differences in residual
variance across groups with SPSS. Or am I on the wrong track generally?
Any help/comments/corrections are greatly appreciated..!
|
Locked
