I am a bit confused on assigning null and alternative hypotheses.

For example in comparing 2 interventions, conventionally, null hypotheses are usually stating that there are no differences in between both interventions. The alternative hypothesis or what we are trying to prove is that there is a difference between both interventions.

However, what if I want to actually prove or my hypothesis in conducting the study is that the interventions are actually equivalent. Won't the conventional null hypothesis become the alternative hypothesis? Will the conventional alternative hypothesis become the null? Will the statistical tests change since we are now assuming that a difference between study arms is true?

  • $\begingroup$ See @gung's answer in particular. $\endgroup$ Oct 7, 2015 at 14:35
  • $\begingroup$ Also see How to test hypothesis of no group differences?, which may offer a suitable alternative. $\endgroup$ Oct 7, 2015 at 15:25
  • $\begingroup$ @Scortchi thanks. I got a few minutes to look and located several somewhat relevant posts but none were better than the two you have already. I'd suggest leaving the closure as is, and I'll delete my initial comment and edit your comment with the second link (you should of course feel free to re-edit it though, since I am putting a few words in your mouth). Feel free to delete this also if you'd like. $\endgroup$
    – Glen_b
    Oct 7, 2015 at 16:11


Browse other questions tagged or ask your own question.