Date:         Tue, 13 Sep 2005 12:03:45 -0500
Reply-To:     "Swank, Paul R" <Paul.R.Swank@UTH.TMC.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Swank, Paul R" <Paul.R.Swank@UTH.TMC.EDU>
Subject:      Re: GLM, comparing levels
Comments: To: Kevin Roland Viel <kviel@EMORY.EDU>
Content-Type: text/plain; charset="us-ascii"

Contrasts without correction will find too many significant results but
doing all pairwise contrasts using a Bonferroni correction will be
overly ocnservation. Under such circumstances, one of the many posttest
comparison proceudres will be more powerful. As an old timer I like the
standard Tukey but some of the others (REQWT) may be a little more
powerful than that.


Paul R. Swank, Ph.D.
Professor, Developmental Pediatrics
Director of Research, Center for Improving the Readiness of Children for
Learning and Education (C.I.R.C.L.E.)
Medical School
UT Health Science Center at Houston

-----Original Message-----
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
Kevin Roland Viel
Sent: Monday, September 12, 2005 6:37 PM
To: SAS-L@LISTSERV.UGA.EDU
Subject: GLM, comparing levels

I have a class variable with 5 levels.  The results of a GLM analysis
suggest that there is a statistical difference in the means between
these levels.  I now am interested in comparing a pair of levels.  It
seems to me that I should perhaps use a CONTRAST statement.  If so,
might anyone suggest the form?

proc glm ;
  class class ;
  model y = class ;
run ;
quit ;

For instance, I would like to compare CLASS = 4 versus CLASS = 5.

Thanks,

Kevin


Kevin Viel
Department of Epidemiology
Rollins School of Public Health
Emory University
Atlanta, GA 30322