stats

Friedman Test

Non-parametric RM-ANOVA — rank matrix heatmap, χ²_r statistic, Kendall W, Conover post-hoc.

Non-parametric alternative to Repeated Measures ANOVA — use when data is ordinal or normality/sphericity assumptions are violated.
Design
S1S2S3S4
Complete data required for all cells — Friedman test cannot proceed with missing values.
Friedman Test Results
χ²_r statistic
df
p-value
Kendall's W
Conclusion α = 0.05
χ²_r = 12/(nk(k+1)) · Σ R_j² − 3n(k+1)  ·  W = χ²_r / (n(k−1))