I wonder if this is a case that it not clearly specified:

Typically, Wilcoxon rank sum test / Mann–Whitney U test is used for independent groups and testing the values of different variables against each other. Typically, Wilcoxon Signed Rank Test is used for dependent groups (same group) and testing the values for the changes in one variable (like in time series study).

But which test is used when I want to test differences for the same group (the same people are asked) for two different variables (non-parametric)? This is not clearly specifiable with the usual definition for these two tests.

For my question the following example: The same teachers are asked about how deeply (or not) they have integrated a specific topic in their courses (Virtual Reality and Data Mining). Here I used a Wilcoxon rank sum test / Mann–Whitney U test with which I want to test if there is a significant difference between the two:

wilcox.test(Bothque$VR_integr, Bothque$DM_integr, exact = FALSE, correct = TRUE, conf.int = TRUE)
  • 1
    $\begingroup$ Can you identify the two answers from a given person (so that the two answers from the same person can be paired)? Or are these answers anonymous so they can't be paired? $\endgroup$
    – Glen_b
    Commented Jul 29, 2022 at 7:55
  • $\begingroup$ Hello Glen, yes they could be paired. These answers are within a group and I can see for each person n (the person themselves are anonymized) which answers (values) are given for which of the two variables. $\endgroup$
    – Martinus
    Commented Jul 30, 2022 at 8:11

1 Answer 1


Assuming I have correctly understood the situation, if you have the data so that it can be paired, you should do so, and use the paired analysis.


Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.