Anavex Life Sciences Corp (AVXL)

[Anshu2]

In that figure (link below) -- which plots the distribution of the mean-change (d) --- you are only trying to determine the probability of going wrong (i.e., the mean-change d whose mean is 1.85 actually becoming < 0; there is no reason to care about d becoming way greater than some positive value). Thus, the area on the LHS is the only area that matters.

So -- I think one-sided is what is correct and makes sense.

Of course, there is another way of looking at p calculation, by looking at the distribution of the t value instead. Therein, it's less obvious whether one-sided or two-sided areas should be looked at.

https://www.investorvillage.com/groups.asp?mb=20622&mn=1054&pt=msg&mid=23759087