Date:         Sat, 1 Nov 1997 19:51:26 -0500
Reply-To:     JOSE FERNANDO DIAZ <jdiaz@WELCHLINK.WELCH.JHU.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         JOSE FERNANDO DIAZ <jdiaz@WELCHLINK.WELCH.JHU.EDU>
Subject:      Re: Poisson regression with autocorrelated errors
Comments: To: Randy Collica <Randy.Collica@HLO.MTS.DEC.COM>
In-Reply-To:  <3450F11F.65AB@hlo.mts.dec.com>
Content-Type: TEXT/PLAIN; charset=US-ASCII

        Dear Randy. Proc Genmod will not take rid of the
autocorrelation. The
covariance structure needs to be defined and this can be done using GEE
and/or GLIMMIXE.SAS The dscale option of GENMOD will however estimate
the overdispersion.
Bye.
 
Jose F. Diaz, MD
Division of Disease Control
Department of International Health
The Johns Hopkins School of Public Health
Phone (home) : 410 685 2155