🔗 Wilcoxon Signed-Rank Test Calculator
Paste before/after pairs to run a Wilcoxon signed-rank test — the rank-based alternative to the paired t-test that doesn't assume normal differences.
Enter the two measurements for each subject in the same order — value 1 in A pairs with value 1 in B.
W statistic
1
z (normal approx.)
2.12
p-value
0.034
W⁺ = 27 · W⁻ = 1 · n = 7 non-zero pairs
Significant difference at α = 0.05
p = 0.034 < 0.05, so you reject the null hypothesis that the paired differences are symmetric around zero — the two conditions differ.
The Wilcoxon signed-rank test is the non-parametric alternative to the paired t-test: it works on the ranks of the paired differences, so it doesn't assume those differences are normally distributed — good for small samples, skewed data or ordinal ratings. Zero differences are dropped; ties share average ranks. This uses the normal approximation with a tie correction (accurate for moderate-to-large n; small samples are better checked against an exact table). 🔒 In your browser.
How the wilcoxon signed-rank test calculator works
The test takes each subject's pair of measurements, computes the differences, drops any that are exactly zero, and ranks the remaining differences by absolute size (average ranks for ties). It sums the ranks of the positive and negative differences (W⁺ and W⁻), takes the smaller as W, and uses the normal approximation with a tie correction to get a z score and p-value. Enter the two conditions in matching order so each row is one subject's pair.
This is the paired-data counterpart to the Mann-Whitney U test (which is for independent groups). Use it in place of a paired t-test when the paired differences aren't normally distributed — before/after measurements, matched subjects, or ordinal ratings. It tests whether the differences are symmetrically distributed around zero. The normal approximation is accurate for moderate-to-large samples; for very small n an exact table is better. Everything runs locally in your browser.
Frequently asked questions
When should I use the Wilcoxon signed-rank test?
For paired or repeated measurements (before/after, matched pairs) when the differences aren't normally distributed. It's the non-parametric alternative to the paired t-test and works on the ranks of the differences.
How is it different from the Mann-Whitney U test?
Wilcoxon signed-rank is for paired data (two measurements on the same subjects); Mann-Whitney U is for two independent groups. Don't mix them up — the pairing is what makes signed-rank appropriate here.
What happens to pairs with no difference?
Pairs whose two values are equal have a zero difference and are dropped from the test, and the sample size n counts only the non-zero differences. Ties among the absolute differences share average ranks.
What does W represent?
W is the smaller of the summed positive-rank total (W⁺) and negative-rank total (W⁻). A large imbalance between W⁺ and W⁻ — a small W — indicates the two conditions differ.
Is my data uploaded?
No — the test runs entirely in your browser and works offline, so your paired data stays on your device.