Date:         Mon, 14 Jan 2008 15:01:41 -0800
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: Pseudo-AIC & Glimmix
In-Reply-To:  <338962.40023.qm@web51002.mail.re2.yahoo.com>
Content-Type: text/plain; charset=iso-8859-1

--- Steve Denham <stevedrd@YAHOO.COM> wrote:
>
> Once more from me-->
> The writers from the television show "The Simpsons" have this covered
> with:
>
>  HOMER: <annoyed grunt>,
>
> which translates as D'oh.
>
> Thank you, Dale.  Once you pointed out the linearization step, I
> realized my error.  Given this, why do you suppose that PROC GLIMMIX
> continues to report likelihood based fit statistics?  I recall that
> when PROC MIXED was first introduced, all of the tests on the
> covariance parameters were automatically presented--but when it was
> pointed out that the tests weren't exactly what they were purported,
> we soon got a fix.  This begs for the same kind of attention.
>
> Steve Denham
> Associate Director, Biostatistics
> MPI Research
>

Steve,

Well, if you have a Gaussian response that you model with the
GLIMMIX procedure, then the likelihood-based statistics are
interpretable.  But how often is one going to model a Gaussian
response using GLIMMIX rather than PROC MIXED?  I would not mind
if all of the likelihood-based statistics were removed from the
GLIMMIX procedure.

On the other hand, with version 9.2, the GLIMMIX procedure is
supposed to incorporate an option for Gaussian quadrature, in
which case we really are attempting to estimate the real response
likelihood function.  In that case, we would want the likelihood-
based statistics to be reported.

Dale


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


      ____________________________________________________________________________________
Be a better friend, newshound, and
know-it-all with Yahoo! Mobile.  Try it now.  http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ