Statistical power calculations: comment.

Power analysis is often used to interpret nonsignificant results (NS). It is a 2-step process: Step 1: Calculate the power of the NS test. Step 2: Use the calculated power to decide whether the NS occurred because (a) the phenomenon was not present or (b) the phenomenon was present but the test had insufficient power to detect it. If power is high, one accepts: (a) the phenomenon was not present. If power is low, one cannot decide between (a) and (b): the NS is not interpretable. This reasoning has minor variations to accommodate the effect size selected for the power analysis. Lenth (2007) argued against using power to interpret NS. Lenth began with an incorrect definition of power.