Statistical Power
Also known as: power analysis, statistical power, test power
The chance your test actually detects a real effect when one exists — low power means you can miss something true and call it nothing.
Statistical Power
Power is the probability a test finds an effect that’s genuinely there. Teams obsess over false positives and ignore the quieter failure: an underpowered test that returns “no difference” when there was one, killing a good idea on bad evidence.
Before I run an experiment I ask what size of effect would matter and whether I have the traffic to see it. If the honest answer is no, the result won’t mean anything — better to know that before spending the weeks than after.