Date: Fri, 14 Mar 2008 14:11:16 -0400
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Mike Zdeb <msz03@ALBANY.EDU>
Subject: Re: segmented regression problem
Hi ... if you are willing to try something other than SAS, you can use the National Cancer
Institute's JOINPOINT software ...
or SEGREG ...
("Joinpoint Goddess", "SEGREG Goddess", hmmmm.... not as universal as "SAS Goddess")
U@Albany School of Public Health
One University Place
Rensselaer, New York 12144-3456
> Help! I am at risk of losing my SAS goddess stature among grad
> I've been asked for SAS help with something out of my area, but which
> seems common enough in the literature that hopefully someone on the
> list can provide some quick code help.
> When analyzing ovarian follicle growth, an important factor is the
> time at which the dominant follicle diverges from the others (follicle
> deviation). This has been done by visual observation of plots of the
> size of largest two follicles (F1 and F2) against time (deviation is
> the time when the lines obviously diverge). A better statistical
> approach is desirable. Segmented regression seems to be apropos.
> "Calculated follicle deviation using segmented regression for modeling
> diameter differences in cattle"
> D. R. Bergfelt, L. H. Segoc, M. A. Bega, and O. J. Ginther
> Theriogenology, Volume 59, Issue 8, 15 April 2003, Pages 1811-1825
> Is basically what I'm after, but I don't know how to get it out of
> PROC REG.
> So with data like:
> F1-F2 (difference in size of F1 and F2)
> How do I get to a calculated time of deviation for each subject?