permutest.cca and means seperation?
Thanks Gavin and Jari, That clears some things up. The results I am getting with capscale make more sense from what I observed in the field then the results I am getting through adonis and betadisper. Still, the tools are great and I can just use both techniques separately, but will watch for the future. Thanks, Michael Wilson
From: Gavin Simpson [gavin.simpson at ucl.ac.uk]
Sent: Thursday, January 20, 2011 4:10 AM
To: Wilson, Michael E
Cc: r-sig-ecology at r-project.org
Subject: Re: [R-sig-eco] permutest.cca and means seperation?
Sent: Thursday, January 20, 2011 4:10 AM
To: Wilson, Michael E
Cc: r-sig-ecology at r-project.org
Subject: Re: [R-sig-eco] permutest.cca and means seperation?
On Wed, 2011-01-19 at 20:23 +0000, Wilson, Michael E wrote: > Hi, > With a CCA object such as outputed by capscale, I see that it is > possible to calculate significance on a class factor with > permutest.cca. > > However, with multiple levels of a factor, is there a means separation > technique for pairwise comparisons like permutest.betadisper but > instead designed to use on centroids identified in a CCA object? betadisper() doesn't do a means separation test - it looks at the dispersions of samples around their respective group centroids and does ANOVA on those dispersions. The analogue of what you discuss is adonis() - MANOVA on any dissimilarity object. It is not quite a direct analogue; capscale and adonis do different things with the non-euclidean parts of the dissimilarities. We vegan developers are discussing what to do about changing capscale so that it works better with the imaginary eigenvectors and is compatible with the way adonis and betadisper handle them. However, this moves the function farther away from Marti Anderson's original CAP procedure. Which reminds me, I owe Jari an email weighing in on these discussions... HTH G Ps - note that it is an object of class "cca" and that is a convenience as we already have tons of code that works on those objects, so we cast the capscale results as an object of that class. > Thanks, Michael Wilson > > > > [[alternative HTML version deleted]] > > _______________________________________________ > R-sig-ecology mailing list > R-sig-ecology at r-project.org > https://stat.ethz.ch/mailman/listinfo/r-sig-ecology -- %~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~% Dr. Gavin Simpson [t] +44 (0)20 7679 0522 ECRC, UCL Geography, [f] +44 (0)20 7679 0565 Pearson Building, [e] gavin.simpsonATNOSPAMucl.ac.uk Gower Street, London [w] http://www.ucl.ac.uk/~ucfagls/ UK. WC1E 6BT. [w] http://www.freshwaters.org.uk %~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%