Date:         Mon, 15 Mar 2010 10:11:17 -0400
Reply-To:     Eric Benchimol <ericbenchimol@ROGERS.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Eric Benchimol <ericbenchimol@ROGERS.COM>
Subject:      Question about Matched Pairs analysis
Content-Type: text/plain; charset=ISO-8859-1

Hello!
I have a relatively simple (I think) question about conducting a
multivariable logistic regression analysis on data which contains
pairs of subjects.  This is not a case-control study, but a survey
with pairs of family members.  I want to assess whether parents are
different from children in their responses (parent vs. patient denoted
by group variable), accounting for other variables.

Each pair of parent and patient share their ID number.

I thought nesting the ID variable within the group variable would
work, but it does not appear to.  I thought of using this program:
The response variable is either 1 or 0.

proc logistic data=whatever;
class group ID response variable1 variable2;
model response = group(id) variable1 variable2 variable3;
run;

Any ideas of how to conduct this analysis using SAS 9.2?

Thanks!