What is funny asbout p<.05 is that while we frequently have a specific hypothesis in mind (e.g. Mean1<Mean2), convention still dictates that we should use a two

What is funny asbout p<.05 is that while we frequently have a specific hypothesis in mind (e.g. Mean1<Mean2), convention still dictates that we should use a two-sided t-test instead of a 1-sided t-test. If I am asked to go to .001, then I will use 1-sided t-tests when I have apriori one-sided hypotheses.