-------- Original Message --------
Subject: Differences between tpsregr and SAS
Date: Thu, 18 Jun 2009 10:43:57 -0700 (PDT)
From: Stephens, Colleen <[email protected]>
To: [email protected]
Hello everyone,
I have a question regarding the differences in calculations made by
tpsregr and SAS.
I tested for common slopes using the example in the Help section. There
were 3 groups with 15 specimens in each group and I used centroid size
as my continuous variable.
Tpsregr gave me one set of statistics. I then performed the same tests
using SAS, but I got a different set of statistics. Some differences
were slight, but for others, the p-values were quite different. In one
instance, tpsregr gave p=0.3 but SAS gave p=0.09. This is almost enough
to draw opposite conclusions.
Is there something special about the multivariate calculations that the
tps series performs that deals specifically with the mathematics behind
morphometrics? Or are these differences arising from slightly different
ways of performing the same calculations in both programs?
Thank you all very much for your time!
Colleen Stephens
Biomathematics Master's Program
School of Biological Sciences
Illinois State University
[email protected]
--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org
