Introduced on day 4
Statistical power
The probability that a test detects an effect of a given size when that effect is real; one minus the Type II error rate.
In detail
Power grows with sample size and effect size and falls with noise. Size experiments for the smallest effect worth acting on, typically at 80% power; a non-significant result from a low-power test says little.