Date:         Mon, 15 Jan 2007 12:43:39 -0600
Reply-To:     Robin High <robinh@UNLSERVE.UNL.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Robin High <robinh@UNLSERVE.UNL.EDU>
Subject:      Re: Proc genmod syntax
Comments: To: kende jan <kendejan@YAHOO.FR>
In-Reply-To:  <443162.45700.qm@web25401.mail.ukl.yahoo.com>
Content-Type: TEXT/PLAIN; charset=US-ASCII

>I have an ordinal dependent variable with 3 classes (Low, moderate, high)
and 4 independent variables.

> I would like to use procedure GENMOD to model this dependent variable.
Which is the best choice to be made for :

- the link function
- the distribution

Hi Jan,

PROC GENMOD only offers cumulative logits as possible links, which may not
be appropriate due to the proportional odds assumption. You should also
look into glogits with PROC LOGISTIC -- it computes a score test for the
POA; GENMOD does not.

Robin High
Univ. of Oregon