>"The concept of power has nothing to do with "degrees of evidence" or vague hypotheses." Yes it does. To properly assess the probability of incorrectly failin

>"The concept of power has nothing to do with "degrees of evidence" or vague hypotheses." Yes it does. To properly assess the probability of incorrectly failing to reject a hypothesis you need to know how likely the data would be under various rival hypotheses. This depends on the precision of the various hypotheses. This is explained by Fisher in my original ref.