Type I & II Errors, Power & Sample Size

This lesson is part of Premium

Type I (α, false positive, 'convict the innocent') vs Type II (β, false negative, miss a real effect). Power = 1−β (conventionally 80–90%) = chance of detecting a real effect. Power ↑ with larger n, larger effect size, lower variability, higher α. Underpowered studies → false negatives. A priori sample-size calculation. SampleSizeDemo. Mnemonic for Type I/II. Exam tip: 'study found no difference' — was it powered to?

Premium unlocks all 40 lessons, every calculator and the timed mock.