Statistical Power

Statistical power is the probability that a test correctly detects a real effect — formally 1 − β, the complement of the Type II error rate. A study with 80% power has an 80% chance of finding an effect of a given size if it truly exists.

Power rises with a larger sample, a bigger true effect, and less noisy data, which is why a power analysis is done before a study to decide how many observations to collect. Skipping it is a classic blunder: an underpowered study not only misses real effects, but the “significant” ones it does find tend to be exaggerated flukes.