Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: trend in ORs across ordered levels of a 3rd variable


From   "Joseph Coveney" <jcoveney@bigplanet.com>
To   "Statalist" <statalist@hsphsun2.harvard.edu>
Subject   RE: st: trend in ORs across ordered levels of a 3rd variable
Date   Wed, 23 Apr 2008 12:37:01 +0900

Maarten buis wrote:

--- "Visintainer, Paul" <PAUL_VISINTAINER@NYMC.EDU> wrote:
Maarten, your model tested the interaction term in the logit.  Am I
right to assume that, in this situation, the Chow test and the logit
test of the interaction term are accomplishing similar goals?
Yes, though I would not use -suest- in this case, because it is more
lines of code, and thus more opportunities for making a typo, your
readers are less likely to be familiar with it, and in this case it
doesn't do anything that can't be done with interactions.

--------------------------------------------------------------------------------

-epitab-'s -mhodds- gives the same test and is even fewer lines of code.

-mhodds married grade3, by(south)-

About Paul's second question:  to test for a linear pattern in the
male-to-female odds ratio over grade.  Peter Lachenbruch mentioned using a
logistic regression model containing a sex-by-grade interaction term.  Paul,
were you able to work this through to do the test that you wanted to?  (Yes,
acknowledging the cautions about interpreting interactions in terms of
probabilities and introducing bias in coefficients when the model omits
important predictors.)

Joseph Coveney


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index