Date:         Tue, 9 Apr 1996 12:04:19 EST
Reply-To:     Ismail Parsa <sip@EPSILON.COM>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Ismail Parsa <sip@EPSILON.COM>
Subject:      Re: Ordinal Regression in GENMOD
Comments: To: lafleurb@ESSEX.UCHSC.EDU

|>  From: Bonnie LaFleur <lafleurb@ESSEX.UCHSC.EDU>
|>  Subject:      Ordinal Regression in GENMOD
 
|>  I am trying find any and all references to doing ordinal regression (on
|>  multinomial reponse) in PROC GENMOD.  Has anyone done this?
 
 
Bonnie:
 
Currently, GENMOD only accepts dependent variables with events/trials format.
If you have an ordinal dependent variable then use proc logistic. The resulting
model is the proportional odds model.  If you would like to model the
non-proprtional odds then consider another software or the neural net
macro (ftp.sas.com, /pub/neural.)
 
Regards,
 
*-----------------------------*
|        Ismail Parsa         |
|   Epsilon Data Management   |
|     50 Cambridge Street     |
|   Burlington MA 01803 USA   |
|                             |
| E-MAIL: sip@epsilon.com     |
| V-MAIL: (617) 273-0250*6734 |
|    FAX: (617) 272-8604      |
|                             |
|  The Usual Caveat Applies   |
*-----------------------------*