|Date: ||Tue, 22 May 2007 10:37:14 -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: proc genmod|
|Content-Type: ||text/plain; charset="us-ascii"|
For 1, you don't need to take the log first as that is the purpose of
the link function. The question is, what is the appropriate
distribution? We don't have enough info to determine that.
Paul R. Swank, Ph.D. Professor
Director of Reseach
Children's Learning Institute
University of Texas Health Science Center-Houston
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
Sent: Tuesday, May 22, 2007 9:56 AM
Subject: proc genmod
using proc genmod with the following syntax
PROC GENMOD data=wtp1 ;
model t2max= dinco t2pain /DIST = gamma LINK= log;
output out = resid resdev=res pred=pr;
some questions arose (I`m a beginner with SAS and generalised linear
1. the dependent variable (all values >=0) is skewed ==> should I use a
log transformed dependent or is this done with the link function above?
2. running the procedure above yielded the following warning:
WARNUNG: Some observations with invalid response values have been
The response was less than or equal to zero for the Gamma
or Inverse Gaussian distributions or less than zero for the
Negative Binomial or Poisson distributions.
What is meant with invalid response values (that the dependent =0)?
Many thanks for your help in advance, Doris