DataTerminology.comNEW
A nonparametric test detecting differences among treatments across multiple blocks in repeated measures or matched designs.
DESCRIPTION
The Friedman test is a rank-based, nonparametric alternative to repeated measures ANOVA for ordinal or non-normal data. It compares multiple treatments across matched subjects or blocks by ranking outcomes within each block. The test statistic assesses whether observed rank sums differ more than expected by chance. Post-hoc procedures can identify which treatments differ. It is robust to violations of normality and often used when assumptions of parametric repeated measures ANOVA are not met.
ALTERNATIVE DEFINITIONS
Wikipedia
The Friedman test is a non-parametric statistical test developed by Milton Friedman for detecting differences in treatments across multiple test attempts.
Emphasizes nonparametric nature and repeated, matched attempts across blocks.
UCLA Institute for Digital Research and Education
The Friedman test is used to determine whether there are differences in the distributions of three or more related groups when the dependent variable is ordinal.
Clarifies application to related groups and ordinal outcomes.
RELATED TERMS
Wilcoxon Signed-Rank Test Nonparametric Statistics Post-hoc Tests Rank-Based Tests Blocking Bonferroni Correction Effect Size Median
COMPETING TERMS
Repeated Measures ANOVA Kruskal–Wallis Test
KNOWLEDGE AREAS
Mathematics Data Science Testing & Validation
TARGET AUDIENCE
Data Scientist Data Analyst BI Specialist Information Analyst
EXTERNAL SOURCES
- Wikipedia - Friedman test
- Penn State Statistics Online - Step-by-step Friedman test lesson
- NIST/SEM e-Handbook - NIST guide on Friedman two-way ANOVA by ranks
- SciPy - Python API for Friedman test
TOP QUESTIONS
When should I use the Friedman Test?
Use it for comparing three or more treatments measured on the same subjects or matched blocks when data are ordinal or non-normal. It replaces repeated measures ANOVA in these cases.
Category: Usage
What assumptions does the Friedman Test make?
It assumes dependent samples across blocks, consistent ranking within blocks, and independence between blocks. It does not require normality of the measurement scale.
Category: Requirements
How is the Friedman Test computed?
Rank data within each block, sum ranks per treatment, and compute a chi-square approximated statistic to test if treatments differ beyond chance.
Category: Process
What post-hoc tests follow a Friedman Test?
Common post-hoc procedures include Nemenyi test or pairwise Wilcoxon signed-rank tests with multiple comparison corrections.
Category: FollowUp
What are common mistakes with the Friedman Test?
Treating independent groups as matched blocks, ignoring ties handling in ranks, and failing to adjust for multiple comparisons can mislead results.
Category: Mistakes
What is the risk of misusing the Friedman Test?
Misapplied matched design or ignoring block effects may inflate Type I error or reduce power, leading to invalid conclusions about treatment differences.
Category: Risk
How does the Friedman Test compare to Kruskal–Wallis?
Friedman is for related samples (blocks), while Kruskal–Wallis is for independent groups. Both are rank-based nonparametric tests.
Category: Comparison
What data types suit the Friedman Test?
Ordinal rankings or continuous measures that violate normality and are collected repeatedly from the same subjects or matched sets are suitable.
Category: Data
Where can I learn the Friedman Test?
Refer to NIST’s e-Handbook, Penn State’s STAT online lessons, and SciPy’s documentation for theory and practical implementation.
Category: GettingStarted