Dear List,
I have some statistical issues with my metacommunity experiment. My
design is the following:
Each metacommunity set consists of five patches, which are connected
with tubes allowing for two dispersal rates (low and high) among these
patches.
Furthermore, I have two different nutrient supply modes: uniform and
heterogeneous.
My response variables are biovolume and species richness.
Now, I would like to test if biovolume (biov) is significant different
among these patches as a function of the dispersal rates (low, high)
and nutrient supply modes (uniform, heterogenous).
Therefore, I used a three-way ANVOA:
model1<-aov(biov~disp*treat*patch, data=loc28), where disp (=dispersal
rate: low and high), treat (=nutrient supply: uniform and
heterogeneous) and patch (= five patches).
I got the comment to use a nested ANOVA, since patches within a
metacommunity are not independent due to connection via dispersal.
So, I used following code: model2<-aov(biov~disp*treat+patch/treat,
data=loc28) but I guess this is not the right way, since disp and
treat are both independent factors and nested in patch.
I would appreciate if someone has an idea to solve this problem or can
recommend another package.
Cheers, Nils
Nils Gülzow
Müllerstr. 21
26384 Wilhelmshaven
Tel.: 04421/3597682
E-mail: nils.guel...@gmx.de
[[alternative HTML version deleted]]
_______________________________________________
R-sig-ecology mailing list
R-sig-ecology@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
