0
$\begingroup$

I'm running a logistic regression with a categorical independent variable. My categorical IV (an ordinal none, low, medium, high) has a pretty uneven distribution: None (4.28%), Low (36.30%, Medium 30.22%, High 29.21%). I have a total sample size of 3,485 so I don't have any sample size issues, but I'm wondering how these uneven groups might affect my estimation. Thanks,

Christina

$\endgroup$
  • $\begingroup$ The group with low observations might have a higher SE, and that's it. $\endgroup$ – user2974951 Aug 7 '19 at 5:53
  • $\begingroup$ Thanks..but the small group is my referent, so would that inflate the SEs in the other categories? $\endgroup$ – Christina Aug 7 '19 at 12:02
  • $\begingroup$ No, but what could inflate the SE of the coefficients is multicollinearity, here it wouldn't matter which group is the reference. $\endgroup$ – user2974951 Aug 7 '19 at 12:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.