alpha diversity interpretation doubt

Good morning to all,
I need an help to understand the alpha diversity analysis of my samples. (Figure attached)

I performed the alpha diversity on a subset representing by different canine breed.
Statistical tests show that a value of Kruskal-Wallis (all groups) are significant but the major values of Kruskal-Wallis (pairwise) were not significant. My doubt is which value should I consider and is there an additional test that I can confirm?
Can anyone therefore explain to me how I should interpret the results?


Hi @rparadiso,

You have two problems here. First, kruskal wallis has issues when there are fewer than 5 items per grooup. So, you have multiple groups where you have two animals, and so the test fails. Second, you have a large number of groups, meaning that you’re paying a pretty large multiple hypothesis correction penalty. My recommendation would be to eeithere see if you can group things more broadly (i.e. terrier, retriever, hound, etc).



Hi @jwdebelius,

Thank for your help! I will try to perform the analysis according your tips.

1 Like

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.