John Gay posted last week on Thursday, May 22nd, 2003:
> Greetings:
>
> Using Stata 7.0, what is the best way to perform
> logistic regression on 1:2 matched (300+) pairs across 3
> clusters (41, 94, 173 pairs respectively) in a case:control
> study. The three independent variables are continuous. Any
> on-line examples? As a novice user of Stata, I can see how to
> handle the matching or the clustering but not both.
>
> Thanks.
> John Gay, Associate Professor, Epidemiology
> Washington State University
--------------------------------------------------------------------------------
Since it's a matched-pair case-control study, it would call for a conditional fixed-effects
logistic regression, -xtlogit , i(pair_id) fe-, or -clogit , group(pair_id)-. With only three
clusters, it doesn't seem as if it would be any benefit to treat cluster as a random effect;
that leaves cluster as a fixed-effect regressor variable that is in dummy (indicator)
variable form. So, how about something like:
-xi: xtlogit outcome continuous1 continuous2 continous3 i.cluster, i(pair_id) fe-
or
-xi: clogit outcome continuous1 continuous2 continuous3 i.cluster, strata(pair_id)-?
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/
