Date:         Mon, 7 Oct 2002 08:58:27 -0700
Reply-To:     Dale McLerran <stringplayer_2@YAHOO.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Dale McLerran <stringplayer_2@YAHOO.COM>
Subject:      Re: Longitudinal data using GENMOD
Comments: To: Silvano Cesar da Costa <silvano@UEL.BR>
In-Reply-To:  <C0E2F977298CD411883100A024B2F510C8CEA7@uel-pdc.uel.br>
Content-Type: text/plain; charset=us-ascii

Silvano,

My guess is that you are overparameterized with the second model.
When you add in the two- and three-way interactions of your factors
with the variable mes, you may end up with some design points at
which there are very few observations.  This can cause the problems
which you report.

Dale


--- Silvano Cesar da Costa <silvano@UEL.BR> wrote:
> Dear Dale and SAS-L,
>
> I have an one-way design with 2 factors, 3 and 4 levels (12
> treatments) with
> 5 repetitions that was followed for 8 months.
>
> My subject is 5*12= 60 Parc.
>
> I am using an correlation struture AR(1) in the genmod. The code
> below works
> very well.
>
> options ls=120;
> proc genmod data=pegas;
> class Enxer Porta Mes Parc;
> model pegas/n = enxer|porta / dist=bin type1 type3 pscale;
> repeated subject=parc / type=ar(1) corrw corrb within=mes modelse;
> run;
>
> But, if I put mes in my model (below), it doesn´t work. Why??? Can
> you help
> me, please???
>
> options ls=120;
> proc genmod data=pegas;
> class Enxer Porta Trat Mes Rep parc;
> model pegas/n = enxer|porta|mes / dist=bin type1 type3 pscale;
> repeated subject=parc / type=ar(1) corrw corrb within=mes modelse;
> run;
>
> The output is:
>
> WARNING: The negative of the Hessian is not positive definite. The
> convergence is questionable.
> WARNING: The procedure is continuing but the validity of the model
> fit is
> questionable.
> WARNING: The specified model did not converge.
> WARNING: Negative of Hessian not positive definite.
> NOTE: The scale parameter was estimated by the square root of
> Pearson's
> Chi-Square/DOF.
> 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.
> NOTE: The scale parameter was held fixed.
> NOTE: The data set WORK.SAIDA has 480 observations and 12 variables.
> NOTE: PROCEDURE GENMOD used:
>       real time           0.25 seconds
>       cpu time            0.23 seconds
>
>
> TIA,
> Silvano


=====
---------------------------------------
Dale McLerran
Fred Hutchinson Cancer Research Center
mailto: dmclerra@fhcrc.org
Ph:  (206) 667-2926
Fax: (206) 667-5977
---------------------------------------

__________________________________________________
Do you Yahoo!?
Faith Hill - Exclusive Performances, Videos & More
http://faith.yahoo.com