Mann-Whitney U Test Calculator
Compare two independent groups by rank, using the normal approximation with a tie correction and an explicit size floor.
Mann-Whitney U Test Calculator
Compare two independent groups by rank, using the normal approximation with a tie correction and an explicit size floor.
What the page prints, and what it refuses
Every figure is printed with the rule or formula that produced it, and any numerical tolerance is stated on the result rather than implied. The page refuses an input the formula does not define and names the value it read; it prints no answer it cannot support.
The calculation runs entirely in this browser tab; nothing is uploaded.
The page ranks two independent groups together and compares their rank sums. With the first group holding the eight lowest of sixteen values, U₁ is 0, z is 3.3082 and the two-tailed p is 0.0009.
It uses the normal approximation, which is why it refuses a group smaller than eight: below that the approximation is poor, and an exact table would be the right tool. The refusal names that reason instead of printing a doubtful p.
What U measures
U counts the pairs where one group's value exceeds the other's. The page computes U₁ from the first group's rank sum, U₂ as n₁n₂ − U₁, and standardises U₁ with a continuity correction of 0.5.
The variance is n₁n₂/12 × ((N + 1) − Σ(t³ − t)/(N(N − 1))), which is the tie-corrected form. The page says whether a tie correction was applied rather than leaving it implicit.
How to use Mann-Whitney U Test Calculator
Enter the first group one value per line, leave a blank line, then enter the second group. Press Run.
The result prints U₁, U₂, z, the two-tailed p, both sizes and the first group's rank sum.
What to Expect
Best for
- Comparing two independent groups on an ordinal scale
- Checking a rank-sum calculation
- Reporting a p when a t-test's assumptions fail
Not ideal for
- Fewer than eight values per group, where an exact test is needed
- Paired observations
- More than two groups, which is the Kruskal–Wallis page
What this tool keeps
- Both groups and their order
What may need cleanup
- Nothing: the figures are final numbers
Common errors
- A group below eight is refused with the reason
- A missing second group is refused with the layout
Example use cases
- Test two groups of eight
- Check a tied comparison
- Compare with a t-test
Sample input
two groups of eight values
Sample output
U₁ = 0, U₂ = 64, z = 3.3082, p = 0.0009
Who this is for
- Researchers with ordinal data
- Students learning rank tests
Frequently asked questions
Why is the minimum eight per group?
The normal approximation to the U distribution is unreliable below about eight per group. The page refuses rather than print a p-value it cannot support.
What is the continuity correction?
U is discrete but the normal curve is continuous, so 0.5 is subtracted from the absolute deviation before dividing. It makes the approximation slightly more conservative.
Are ties allowed?
Yes. Tied values share average ranks and the variance is corrected for them; the result line says whether a correction was applied.
Build custom tools for your store
Join the SDK program to create useful ecommerce utilities for your store.