AMOVA + unequal sample size


I would like to use AMOVA to describe the dissimilarity between the communities of 2 different treatment groups (i.e. infected/non-infected). However, my sample size is uneven (inf=5, non-inf=16).

How well does AMOVA tolerate such a big difference in sample size? In the original paper, it does state that unequal sample sizes are tolerated but I amn’t sure to what extent.

Many thanks,


You should be good to go - it’s a non-parametric test that uses a randomization step.


Many thanks for this, Pat!