Melanie,

I can't tell you what SAS will or will not do, but I do know that 
NMDS does not provide proportions of variance explained by axes in 
the same way that PCA does.

One thing you can do however, (which PC-Ord will do for you, or if 
you have enough time, you could do on a spreadsheet), is to simply 
run a regression of the distances in ordination space for each axis 
against the original distances (which either you provided to SAS or 
were calculated for you in Proc MDS; e.g., Sorensen distances, 
Euclidean distances, etc.). You could then calculate coefficients of 
determination (r2) for each axis. I doubt that SAS will provide the 
distances in ordination space for you (although it might calculate 
the original distances).

McCune and Grace's book Analysis of Ecological Communities gives a 
pretty good (probably better) explanation of what I'm talking about.

Good luck,
Steve Brewer


>Does anyone know how to get SAS to output the proportion of variance
>explained by each axis when using Proc MDS?
>
>Thank you.
>Melanie


-- 
Department of Biology
PO Box 1848
University of Mississippi
University, Mississippi 38677-1848

Brewer web page - http://home.olemiss.edu/~jbrewer/

FAX - 662-915-5144
Phone - 662-915-1077

Reply via email to