stats

Mann-Whitney U Test

Nonparametric two-group test — U statistic, Z approximation, p-value, effect size r, rank-annotated dot plots.

Use this instead of a T-test when your data is ordinal or not normally distributed.

Input — Group Data

Hypothesis Tail

Group A (comma or space separated)Group B (comma or space separated)

Results

n₁ (Group A) / n₂ (Group B)—

Rank sum R₁ / R₂—

U₁—

U₂—

U statistic (min of U₁, U₂)—

Z (normal approx)—

p-value—

Effect size r = |Z|/√N—

Conclusion α = 0.05—

Conclusion α = 0.01—

U₁ = n₁n₂ + n₁(n₁+1)/2 − R₁ · ties → avg ranks
Z = (U − n₁n₂/2) / √(n₁n₂(N+1)/12) with continuity correction