Date:         Tue, 16 Sep 2008 09:32:24 -0400
Reply-To:     Kevin Viel <citam.sasl@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Kevin Viel <citam.sasl@GMAIL.COM>
Subject:      Re: Proc genmod repeated measurements

On Mon, 15 Sep 2008 16:39:01 -0700, Ying Yang <yinyang@UCDAVIS.EDU> wrote:

>I have a dataset with 332 observations. Among 332 observations, there
>are 26 sets of brothers.  In order to take into account of related
>measures for brothers, I  use proc genmod with repeated measurements.
>The SAS code and error messages listed below. Do you have any
>suggestions for analyzing this data regarding the error messages below?
>Any help is highly appreciated. Thanks,
>
>
>proc genmod data=aa descending;
>class ID group;
>model outcome = group  / dist=bin link=logit type3 wald;
>repeated subject=ID/corrw;
>run;
>
>WARNING: The generalized Hessian matrix is not positive definite.
>Iteration will be terminated.
>
>ERROR: Error in parameter estimate covariance computation.
>ERROR: Error in estimation routine.


Why do you have the REPEATED statement?  It seems that you might have
(unbalanced) multiple measurement: 12 X 26 = 312 > 332.  Might you also
want a RANDOM statement?  Are any of the sibling MZ (monozygotic, i.e.
identical)?  Could you describe your study and its goals, including
sampling design?

HTH,

Kevin