Re: st: Testing equality of odds ratios outside of logistic regression

["Michael I. Lichter" <mlichter@buffalo.edu>]

---
I want to thank Michael Mitchell, Steve Samuels, and Tim Mak for their 
helpful responses to my query of 10 days ago (sorry for this very late 
follow-up!).

Michael's suggestion of using -suest- was the most immediately helpful. 
I've always been a bit leery of using -suest- and -sureg- because I 
don't have any textbooks that explain the underpinnings of the method. 
Am I correct, by the way, that in lieu of a single omnibus test, I can 
test each pair of variables individually after -suest- and use -test, 
accumulate- to get a single p-value?

Steve's suggestions re. logistic regression with interactions or a 
homogeneity test within -cc- appear to apply only to mutually exclusive 
categories of a single nominal or ordinal variable. My variables are not 
mutually exclusive.

Suggestions from Tim and Steve regarding prediction, validation, 
-roccomp-, etc., will be helpful for future efforts. I'll probably 
revisit this in a month or so; any reading recommendations would be 
appreciated.

Thanks again.

Michael


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/