I have been searching replication studies and I found that all of them use two-tailed tests. For example:

Open Science Collaboration. (2015). Estimating the reproducibility of psychological science. Science, 349(6251), aac4716.

I am wondering why one-tailed tests using the direction of the original study are not used instead?


1 Answer 1


I guess that your logic in here is that assuming that we know that the effect was negative (or positive), let's use one-tailed test, however the assumption is wrong, because you are conducting the replication study to check if the results hold, so you cannot take the direction for granted. Also, the choice between one-tailed and two-tailed test depends on the hypothesis you want to test, not what you've seen in the data. Either your hypothesis is directional, or not. You are testing against the assumption what the result could be, not against what you think it wouldn't be.

Moreover, it is easier to get significant $p$-value with one-tailed, as compared to two tailed test. If you choose the easier to pass test to verify the initial results, then you are not helping the cause. Obviously, if the easier test wouldn't be passed, this would be suspect, but if it is passed, this does not prove that the harder test would be passed.


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.