I am considering using PERMANOVA (vegan: adonis) and Multivariate Analysis
of Dispersion (DISPER; vegan: betadisper) to test whether differences in
soil chemical composition occur between two stand types.  My reasoning for
using these techniques rather than parametric MANOVA is that, if PERMANOVA
shows differences between stands, I would like to know if those differences
are driven (at least partly) by variability of soil characteristics between
groups, using DISPER.  My idea is to center and scale the soils data (they
vary in units of measure) and then just proceed with functions adonis and
betadisper in vegan using euclidiean distances.

Does this sound like an appropriate course of action?

****************************************
Chris Habeck
Remnant Expansion Project
Postdoctoral Research Associate
Department of Plant Biology
Michigan State University
Website: http://habeck-ecology.weebly.com/index.html

        [[alternative HTML version deleted]]

_______________________________________________
R-sig-ecology mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-ecology

Reply via email to