I'm trying to figure out what the default alternative hypothesis of this test means.

If my significance threshold is 0.05, is it incorrect to say only p-values < 0.05 are significant or are values > 0.95 also significant?

I just want to say the distribution of vector u is significantly different than vector v regardless of the direction (if this makes sense).

https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.mannwhitneyu.htmlenter image description here

  • 1
    $\begingroup$ A p-value > 0.95 is not significant. The two-sided nature of the test is taken into account by the "alternative" parameter, so that you just need to worry about a single 0.05 threshold. ... Also, the hypothesis for the Mann Whitney test in the question isn't stated quite correctly. $\endgroup$ – Sal Mangiafico Sep 12 '17 at 1:57
  • $\begingroup$ mostly duplicated here stats.stackexchange.com/questions/259112/… $\endgroup$ – Glen_b Sep 12 '17 at 2:27
  • $\begingroup$ While the link covers most of the question (which is why I initially closed as duplicate), I've reopened as the last aspect of the question isn't dealt with there. I'd suggest editing to focus more explicitly just on that aspect if possible $\endgroup$ – Glen_b Sep 12 '17 at 2:50

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.