|Date: ||Thu, 15 Jan 2004 11:40:35 -0600|
|Reply-To: ||Paul R Swank <Paul.R.Swank@UTH.TMC.EDU>|
|Sender: ||"SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>|
|From: ||Paul R Swank <Paul.R.Swank@UTH.TMC.EDU>|
|Subject: ||Re: Regression|
|Content-Type: ||text/plain; charset="US-ASCII"|
Also, how was gender coded? That will make a difference for the sign.
Paul R. Swank, Ph.D.
Professor, Developmental Pediatrics
UT Health Science Center at Houston
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
Sent: Thursday, January 15, 2004 11:12 AM
Thanks for all the responses....
1) I did include the main effects as well as the interaction term in the
model. I first centered the predictor variable and used that new variable
in the model to help reduce colinearity.
2) There are nearly twice as many women as men in the sample. Would this
effect the standard error estimates (DF's)?