For example, how would you decide between which tailed test to do in the following experiment:

Ethanol dilutions were made with different proportions and Daphnias were submerged in them to collect their heart rates. The Daphnias were also submerged into beakers with different temperatures.

My null hypothesis was that the Daphnia will not be affected by the changing variables and if they are, it is due to a purely kinesis response (caused by chance).

My alternative hypothesis was that Ethanol would decrease the heart rates of the Daphnias as the dilution concentrations increase. I assumed this because Ethanol is a respiratory and nervous system depressant that acts as an anesthetic and potentially leads to death after an overdose. I also assumed that the Daphnias' heart rates would fall and become steadier in lower temperatures because they usually inhabit cooler, freshwater habitats.

  • 1
    $\begingroup$ The body text of your question is very unclear, but in respect of your title -- it depends on whether your alternative hypothesis is one sided or two sided ... and we can't really tell you what to hypothesize. $\endgroup$ – Glen_b -Reinstate Monica Mar 7 '15 at 7:24
  • 1
    $\begingroup$ I must agree with the first comment by Glen_b. However, even if we cannot tell you what to hypothesize, it could be beneficial for you to learn more about both options. See, for example, here (sections "When is a one-tailed test appropriate?" and "When is a one-tailed test NOT appropriate?"). By the way, your question should be quite popular in general so there is quite likely a lot of material about it. Introductory statistics textbooks should have some discussion on the topic. $\endgroup$ – Richard Hardy Mar 7 '15 at 14:58
  • 1
    $\begingroup$ Please do not vandalize your question. $\endgroup$ – gung - Reinstate Monica Mar 8 '15 at 3:07
  • 2
    $\begingroup$ It's utterly pointless to keep trying to vandalize your question (what is it, four attempts now?). It will simply be rolled back to a valid version every time. If there's some very important issue, you can flag your post and ask a moderator about whether it can be deleted but unless it's something extremely important, I really don't imagine that it will be deleted. Why do you want to remove it? $\endgroup$ – Glen_b -Reinstate Monica Mar 10 '15 at 3:37

If your result was significant in either direction, would this be note-worthy? In other words, if the data shows that Ethanol actually increases, rather than decreases, the heart rates of the Daphnias as the dilution concentrations increases would this be an interesting result worthy of being noted? If so, do a two tail test.

If it is not possible that Ethanol actually increases the heart rates of the Daphnias, based on past research in the field or common knowledge, do a one-tail test. Of course, due to random chance, the data could suggest Ethanol increases the heart rates. In this situation, we can chalk this up to error hence not note-worthy.

  • 4
    $\begingroup$ @OminousOwl -- please don't vandalize the answer, either. Your attempted edits will simply be rejected on review, and even if they somehow ever get through that review stage, the edit would then be rolled back. You don't get to vandalize posts on the site, even your own. You can delete your own question, right up until the point that it's been answered -- but then it's not yours to delete any more, since it has someone else's content in it. $\endgroup$ – Glen_b -Reinstate Monica Mar 10 '15 at 3:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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