Date:         Fri, 3 Oct 2003 21:12:12 -0700
Reply-To:     Sharon <findingjobs@HOTMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Sharon <findingjobs@HOTMAIL.COM>
Organization: http://groups.google.com
Subject:      Generalized estimating equations
Content-Type: text/plain; charset=ISO-8859-1

I'm using GEE (exchangeable correlation matrix) for my analysis
because my data are correlated. In PROC LOGISTIC (SAS), I can specify
build a model using a stepwise, forward or backward procedure.
However, since these procedures use the maximum likelihood, what is
the equivalent model building tools for a GEE,using PROC GENMOD, since
GEE uses a quasi-likelihood approach? Thanks.