Date:         Thu, 29 Nov 2007 17:47:22 -0800
Reply-To:     Robin High <robinh@UOREGON.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Robin High <robinh@UOREGON.EDU>
Subject:      Re: logistic regression output,
              est. & SE: JMP (made by SAS) vs. Stata
Comments: To: Susan Lingle <susan.lingle@ULETH.CA>
In-Reply-To:  A<474F65BC.4000500@uleth.ca>
Content-Type: text/plain; charset="us-ascii"

Susan,

I don't have JMP or Stata, though the symptoms you describe indicate
that Stata applies effect coding (-1,1) to class variables and JMP
applies GLM coding (0,1). Check the documentation to see which type of
coding is used in each.

Robin High
University of Oregon


-----Original Message-----
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
Susan Lingle
Sent: Thursday, November 29, 2007 5:22 PM
To: SAS-L@LISTSERV.UGA.EDU
Subject: logistic regression output, est. & SE: JMP (made by SAS) vs.
Stata

This question does not involve SAS, but it involves JMP, which is made
by SAS and many SAS people are familiar with JMP (well, I hope so :)).
Does anyone know why I am obtaining parameter estimates and SE for a
logistic regression that are exactly twice as high in Stata than in JMP?
Values for the overall models (LR and p-values) are identical. I haven't
found any details to shed light on how these are calculated in the two
programs.

Susan