Compositional Data Analysis?

[deleted_user]

I am trying to determine if there is a difference between two sets of data that are compositional (add up to 100%) and therefore not independent.

Does anyone know and have an example of how to do this using JMP or SAS? I believe you have to use MANOVA to analyze the two sets of data in which variables are represented as proportions ("logratio analysis of compositions"). Thank-you!

[Susan]

Barry Moser at LSU has an example program posted at

http://www.stat.lsu.edu/faculty/moser/exst7037/exst7037.html

Scroll about about half-way down. In the second GLM call, you'll have to change

MNAMES="Y3"

to

MNAMES=Y3

(at least in 9.2)

Have fun!

[Flip]

Hey he was my graduate advisor 18 years ago.