The default assumption in statistical testing that there is no real effect or pattern, used as a baseline to test against.