Interpreting one-tail P values

One-tail P values (also called one-sided P values) can only be interpreted if you predicted the direction of the effect before collecting data. If you are looking at differences, predict an increase or predict a decrease. If you are looking at correlation, predict a positive correlation or predict a negative correlation. If you are looking at a relative risk, predict a relative risk greater than 1.0 (exposure increases risk) or predict a relative risk less than 1.0 (exposure reduces risk). 

One tail P values reported by Prism are only correct if you correctly predicted the direction of the effect. 

If the actual effect was in a direction opposite to your prediction, then the correct one-tail P value equals 1.0 minus the reported one-tail P value. This correct P value in this case (when your prediction was backwards) is always greater than 0.50.