I've got a set of data points with "scores" that belong to a series of subgroups (not necessarily disjoint). That is, I may have a point that belongs to groups A, B, and D, and then another point that belongs to B, C, and E, and yet another that belongs to group A, D, and F.

Within each subgroup (say, subgroup A), points belong to one of five conditions. These are disjoint (no point belongs to two conditions). Initially, I was running a Tukey-Kramer HSD test to find whether condition 1 scored higher than condition 2 and so on.

However, I have 15 subgroups. I understand that the Tukey-Kramer HSD method corrects for the multiple comparisons problem within each group, but if I have multiple sets of tests, then what kind of correction can I make to account for this? I believe the Bonferroni would be incredibly conservative (I'd have 1500 individual tests in all), and I don't know how I might be able to apply the Benjamini-Hochberg in this scenario.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.