Experimentation
Statistical power
Also calledPower
The chance that a test detects a real effect when there is one. Tests are usually planned for 80 per cent power.
Power is the other side of significance. Significance protects you from false positives; power protects you from false negatives, where a change that really works is dismissed because the test was too small.
An underpowered test is worse than it looks: when it does find a winner, the measured effect tends to be exaggerated.