Nonparametric
Wilcoxon signed-rank test
Compares two measurements on one group by ranks.
Variables
- First measurementContinuous · Ordinal — 1
- Second measurementContinuous · Ordinal — 1
Options
- Alternative hypothesisTwo-sided · One-sided (greater) · One-sided (less)
- Continuity correction
What it reports
The rank-based counterpart for paired samples: T, Z, p and an effect size.
scipy+formula · Open the accuracy report
When does this apply?
The same design as the paired-samples t-test — one group’s measurements in two separate columns, one row per observation — but the comparison is by the ranks of the differences, so the measurements may also be ordinal. The output is T, Z, p, and the rank-biserial correlation with a confidence interval.