All tools

Friedman test

The Friedman test compares three or more measurements of the same people, for example at the start, middle and end of a program, or the rating each person gives three products. It is the non-parametric alternative to repeated-measures ANOVA and works with Likert scales. Paste your data with one row per person and one column per measurement.

You can copy the columns from Excel, SPSS or Google Sheets. The first row can hold each measurement’s name.

The calculation runs in your browser. We don’t store what you type.

How it’s calculated

Within each row (each person) scores become ranks: 1 for the lowest. Ties get the average rank.

Each column’s ranks are summed (Rⱼ). χ² = 12 / (n k (k + 1)) × Σ Rⱼ² − 3 n (k + 1), with n people and k measurements.

With ties, χ² is divided by 1 − Σ(t³ − t) / (n (k³ − k)), as in SPSS.

The p-value comes from the chi-square distribution with k − 1 degrees of freedom.

The effect size is Kendall’s W: W = χ² / (n (k − 1)). It ranges from 0 to 1.

Pairwise comparisons divide the difference in mean ranks by √(k (k + 1) / 6n), with the Bonferroni correction, as in SPSS’s model view.

How to read the result

  • p < .05: at least one measurement differs from the others. Check the pairwise comparisons to see which.
  • Report each measurement’s median and mean rank, not the means.
  • Kendall’s W as effect size: 0.1 small, 0.3 medium and 0.5 large.

In the glossary: Friedman test · p-value · Effect size · Parametric and non-parametric tests

Next stepEffect size calculatorAPA style asks for the effect size next to every test: compute it from the result you just got.Step-by-step guideFriedman test step by step: when to use it, how to compute it and how to report itFriedman test step by step: when to use it, worked example, pairwise comparisons, Kendall’s W and how to report it in APA 7.

Frequently asked questions

Does it match SPSS?
Yes. Chi-square, degrees of freedom, p-value and mean ranks match Analyze > Nonparametric Tests > Legacy Dialogs > K Related Samples. Pairwise comparisons match the “Adj. Sig.” in the model view.
Friedman or repeated-measures ANOVA?
If the measurements are normal, repeated-measures ANOVA has more power. If not, or the scale is ordinal, use Friedman.
What if I only have two measurements?
With two measurements (before and after) use the Wilcoxon signed-rank test or the paired t-test.
Can I paste data with headers?
Yes. If the first row has names (Start, Middle, End), they are used to label the results.

Know what you want to ask?

Build the survey in SurvIQ, share it by link, WhatsApp or QR code, and watch results come in live. The Free plan is free forever.

Create a free survey

More tools