I'm performing Friedman test to see if there is a difference between 8 treatments, then if it is significant I'll do Wilcoxon signed rank test to see which pair is significant. I wonder if I should adjust this for multiplicity or this isn't needed because it will be performed only if Friedman test is significant at 5% level. Please help.

  • $\begingroup$ Please note that Friedman analysis is not an extension of Wilcoxon from 2 to k groups. You should use either Quade analysis or perform the Sign test after Friedman. $\endgroup$
    – ttnphns
    Commented Jan 2, 2023 at 12:03
  • $\begingroup$ What do you mean by "multiplicity"? Is it about the multiple comparison correction? $\endgroup$
    – ttnphns
    Commented Jan 2, 2023 at 12:05
  • $\begingroup$ Thank you for your response. Yes, by multiplicity I meant if I should adjust alpha level due to multiple comparisons performed using wilcoxon test for each pair $\endgroup$
    – user365345
    Commented Jan 2, 2023 at 13:05


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.