reduced major axis regression (RMA)

[Phil Townsend]

Cahn anyone provide advice on how to do RMA (reduced major axis)
regression in SAS ?

THanks....

--
Phil Townsend
Associate Professor
Department of Forest Ecology and Management
University of Wisconsin-Madison
1630 Linden Drive, Russell Labs
Madison, WI 53706

phone: [phone] fax: [phone]
email: ***@wisc.edu

[David L Cassell]

Post by Phil Townsend
Cahn anyone provide advice on how to do RMA (reduced major axis)
regression in SAS ?
THanks....
--
Phil Townsend
Associate Professor
Department of Forest Ecology and Management
University of Wisconsin-Madison
1630 Linden Drive, Russell Labs
Madison, WI 53706
phone: [phone] fax: [phone]

First, let me suggest that RMA is usually not the best thing to do.
Do you have a simple linear Y vs. X relationship but with additive
measurement error on X? And no other regressors? And no
other data problems, like outliers, leverage points, etc.? If so,
and your measurement error is fairly substantial in size compared to
the variability of Y and X, then RMA may be a good choice. If
not, then perhaps there are other methods which would be
better.

In the meantime, RMA is more typically called errors-in-variables
regression in the stat biz. SAS does this, using one of two procs.

I like doing this in PROC NLP (which is in SAS/OR) just because
I'm eccentric.


proc nlp data=YourData;
min dist;
parms b1=1, b0=1;
dist=(y - (b0 + b1*x))**2 / (1 + b1*b1);
run;


That's it. It is explicitly finding the minimum perpendicular
distance to the regression line. That's RMA.

Or you can do this using PROC CALIS (which is in SAS/STAT).


If you want to do bootstrapping of your data to get CIs,
you can use some of my bootstrap code from the SAS-L
archives to build a bootstrap data set and then use
by-processing to feed the bootstrap data set through
either proc.

HTH,
David
--
David L. Cassell
mathematical statistician
Design Pathways
3115 NW Norwood Pl.
Corvallis OR 97330

_________________________________________________________________
View Athlete’s Collections with Live Search
http://sportmaps.live.com/index.html?source=hmemailtaglinenov06&FORM=MGAC01