[R-sig-eco] post-hoc test for adonis PERMANOVA

[Jari Oksanen]

There is a parametric post hoc test for betadisper(), but there is no post hoc test for adonis(). These two functions are different and study different null models. If anybody figures out how to do post hoc test for adonis() we may incorporate submitted code into vegan (though personally I think that the whole idea of post hoc tests is weird, to put it mildly).

Cheers, Jari Oksanen

Sent from my iPad

On 25.11.2012, at 0.54, "Andre Rovai" <asrovai at gmail.com> wrote:

> Dear friends,
> 
> I have been recently introduced to R and am dedicating myself to go deeper
> on it. It can be tricky sometimes for a rookie, though!
> 
> I ran a PERMANOVA (vegan package) using the adonis() function and tried to
> go deeper in the analysis by running a post-hoc test (betadisper()). I do
> appreciate if someone more familiarized with codes related to the post-hoc
> analysis could validate the way I am headed to. Here are the codes I used:
> 
> # permanova
> fl.sem.transf.bray <- vegdist(fl, method="bray")
> flfat <- read.table("flfator2.txt", header=T, row.names=1)
> adonis(fl.sem.transf.bray ~ loc*trat, data=flfat, strata=flfat$trat,
> permutations=999) # there was significant difference for the interaction
> 
> # post-hoc test
> phoc <- with(flfat, betadisper(fl.sem.transf.bray, trat))
> TukeyHSD(phoc)
> 
> Thank you!
> Andre Rovai
> 
>    [[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