Date:         Wed, 28 Nov 2007 12:10:00 -0500
Reply-To:     Wensui Liu <liuwensui@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Wensui Liu <liuwensui@GMAIL.COM>
Subject:      Re: Poisson model and NB model
Comments: To: jn mao <jn_mao@yahoo.com>
In-Reply-To:  <303275.48840.qm@web32714.mail.mud.yahoo.com>
Content-Type: text/plain; charset=ISO-8859-1

jn,
as far as I understand, the models you mentioned are nested. if so,
then take a look at LR test.

On Nov 28, 2007 11:16 AM, jn mao <jn_mao@yahoo.com> wrote:
> Hello SAS-Ls,
>
>  Could someone give me some advice about how to choose adjusted Poisson model and NB model?
>  I'm analyzing an over-dispersed data, the NB model shown one variable is not significant, but when I used scale deviance adjust poisson model, that variable becomes significant. I'm not sure which model I should use? why the adjusted poisson model does not consist wit£è¡¡
>  NB model?
>
>  Jane
>
>
> ---------------------------------
> Get easy, one-click access to your favorites.  Make Yahoo! your homepage.
>



--
===============================
WenSui Liu
Statistical Project Manager
ChoicePoint Precision Marketing
(http://spaces.msn.com/statcompute/blog)
===============================