How to use
- Paste independent Group A and Group B values in separate boxes.
- Choose whether the directional question is Group B > Group A, Group B < Group A, or two-sided.
- Read the p-value together with both U statistics, rank sums, and whether ties forced the page to use a normal approximation.
Wave 36 independent nonparametric entry
Read an independent rank shift
The share URL stores only the selected tail and decimal places.
How to choose between nearby pages
Use two-sample t-test when mean-based independent-group comparison is fine. Use Wilcoxon signed-rank when the same items are paired before and after. Use this page when two groups are independent and the question is a rank-based shift rather than a paired change.
FAQ
What does this page calculate?
It runs a Mann-Whitney U test on two independent groups and reports the selected p-value, both U statistics, rank sums, and whether the page used an exact or asymptotic path.
When should I use Mann-Whitney U instead of a two-sample t-test?
Use Mann-Whitney U when two groups are independent but the data are ordinal, skewed, or too non-normal for a comfortable mean-based two-sample t-test read.
Why is the page using a normal approximation?
If pooled values tie, the exact U distribution is not used here. The page switches to a normal approximation and warns you when that happens.
Does the share URL include my group values?
No. The share URL stores only light settings such as the alternative and decimal places. Your group values stay in your browser.