MegaMaester

Statistics for Everyday Life

Type I vs Type II Error

In any test or decision under uncertainty, two opposite mistakes are possible. Statisticians call them Type I and Type II errors — the false positive and the false negative.

AspectType I errorType II error
Also known asFalse positiveFalse negative
What happensDetecting an effect that isn’t realMissing an effect that is real
In a court analogyConvicting an innocent personAcquitting a guilty person
Medical test analogyHealthy person flagged as sickSick person told they’re healthy
Controlled byThe significance threshold (e.g., 5%)Sample size and effect size (statistical power)

When to use type i error

Worry most about Type I errors when a false alarm is costly — approving a useless drug, or convicting the innocent — and set a stricter threshold.

When to use type ii error

Worry most about Type II errors when missing a real effect is costly — failing to catch a disease, or overlooking a real safety risk — and use a larger sample for more power.

Frequently asked questions

How do I remember which is which?
Type I = false positive (you saw something that wasn’t there). Type II = false negative (you missed something that was). One memory aid: "Type I, you cry wolf; Type II, you miss the wolf."
Can you avoid both at once?
There’s a trade-off: making a test stricter to reduce false positives tends to increase false negatives, and vice versa. The main way to reduce both together is to gather more data (a larger sample), which increases statistical power.
What is statistical power?
Power is the probability of correctly detecting a real effect — that is, of avoiding a Type II error. It rises with larger samples, bigger true effects, and less noisy data.