follow-up on "Proper way to perform many pair-wise comparisons?"

[mikerobeson]

Moved this thread (joey711/phyloseq#330) here as requested.

[joey711]

Is there a topic?

[mikerobeson]

Yep, we are concerned about those multiple-multiple comparisons too. We are just doing some simple data exploration at the moment. But, as mentioned before, depending on how I carry-out these pair-wise comparisons via phyloseq (linked in the previous thread post above) I get quite different results. I am trying to get a handle on how those different approaches (i.e. procedure 1 vs procedure 2.) affect the interpretation of the results. And which of the two is the generally (technically) preferred way to perform pair-wise comparisons. Any insights?