Should significance decisions in Dunn's test be based on the adjusted p-value or the raw p-value?
I am performing post-hoc pairwise comparisons after a significant Kruskal Wallis test using Dunn’s test.
The output of the function reports both:
Unadjusted p-values
Adjusted p-values (Benjamini–Hochberg)
For determining statistical significance, should the decision be based on the adjusted p-values or the raw p-values?