Kendall rank correlation coefficient¶

A rank-association statistic based on the excess of concordant over discordant observation pairs.

Version: v2 · 2026-09-19 · History
Domain-specific #: 5185
Domain group: Formal Sciences
Origin domain: Experimental Design & Statistics
Subdomain: Nonparametric Statistics → Experimental Design & Statistics
Aliases: Kendall's tau, Kendall tau coefficient

Core Idea¶

Tau-a assumes no tie correction, while tau-b and tau-c normalize for ties or rectangular tables; sampling, pair independence and exact or asymptotic inference must match the chosen form. Every pair of observations is classified by whether the two variable orderings agree, disagree or tie, and a normalized signed difference measures ordinal association. The abstraction is therefore identified by a declared carrier, a transformation or constraint over that carrier, and an invariant that tells an analyst whether the named structure is genuinely present.

How would you explain it like I'm…

Do We Agree on Pairs?

Two friends each line up the same toys from favorite to least favorite. Pick any two toys and see if both friends put them in the same order. If most pairs match, the friends agree a lot; if most pairs are flipped, they disagree a lot.

Pair-by-Pair Agreement Score

Suppose you have two ways of ranking the same things, like how fast kids run and how high they jump. Look at every pair of kids: if the faster one also jumps higher, that pair 'agrees'; if the faster one jumps lower, it 'disagrees'; if two are equal, it's a tie. The Kendall rank correlation coefficient counts agreeing pairs minus disagreeing pairs and scales the result, so it shows how strongly the two orders go together. Different versions handle ties in different ways.

Concordant-Pair Rank Correlation

The Kendall rank correlation coefficient (tau) measures how much two orderings of the same items agree. Every pair of observations is classified as concordant (both variables order the pair the same way), discordant (opposite ways), or tied. Tau is the difference between concordant and discordant counts, divided by a normalizing denominator, so it ranges from strong disagreement to strong agreement. Unlike Pearson correlation, it uses only order, not the actual distances between values. Tau-a makes no correction for ties, while tau-b and tau-c adjust for ties or for tables that aren't square, so you must say which version you used.

The Kendall rank correlation coefficient is a nonparametric measure of ordinal association between two variables measured on the same sampling units. Each pair of observations is classified as concordant, discordant, or tied according to whether the two variables' orderings agree, disagree, or tie, and tau is a normalized signed difference of these counts. Tau-a uses no tie correction; tau-b corrects the denominator for ties; tau-c is adapted to rectangular contingency tables. The variant fixes the denominator and hence the attainable range. Inference against a null of no association uses either an exact distribution or an asymptotic variance, and this must match the chosen variant, the sampling design, and the assumption that pairs come from independent observations. A proper report states the paired units, the ordering of each variable, missing-data handling, the tie definitions, the variant, and the uncertainty.

Scope of Application¶

Kendall rank correlation coefficient belongs to nonparametric statistics and is useful where the analyst can specify the typed nonparametric statistics carrier, including its objects, relations, parameters, conventions, evidence, boundary cases, and comparison targets, then evaluate the paired observations and sampling unit, ordinal variables and ordering, missingness, concordant discordant and tied-pair definitions, tau variant and denominator, estimate range, null hypothesis, variance or exact distribution and uncertainty are explicit. The scope is broad within that domain but bounded by the need for the paired observations and sampling unit, ordinal variables and ordering, missingness, concordant discordant and tied-pair definitions, tau variant and denominator, estimate range, null hypothesis, variance or exact distribution and uncertainty are explicit.

Clarity¶

The abstraction clarifies a crowded vocabulary by making the paired observations and sampling unit, ordinal variables and ordering, missingness, concordant discordant and tied-pair definitions, tau variant and denominator, estimate range, null hypothesis, variance or exact distribution and uncertainty are explicit the center of the account. A claim should name the carrier, the governing operation or relation, the applicable assumptions, and the recognition test.

Manages Complexity¶

Without the abstraction, an analyst must reason directly over many local details: the carrier roles, admissibility assumptions, competing conventions, derived invariants, boundary cases, and proof or validation obligations specific to Kendall rank correlation coefficient. Kendall rank correlation coefficient compresses them into the roles in the structural signature. That compression permits comparison across instances without erasing the variables that determine validity. It also exposes which details may be varied safely and which are constitutive.

Abstract Reasoning¶

Identify the carrier. State what the elements, states, objects, or observations are: the typed nonparametric statistics carrier, including its objects, relations, parameters, conventions, evidence, boundary cases, and comparison targets. Reject examples whose alleged carrier belongs to a different problem.
Lock the constitutive rule. Express the paired observations and sampling unit, ordinal variables and ordering, missingness, concordant discordant and tied-pair definitions, tau variant and denominator, estimate range, null hypothesis, variance or exact distribution and uncertainty are explicit independently of one notation or implementation.

Knowledge Transfer¶

Knowledge transfers strongly among subfields of nonparametric statistics because they reuse the typed nonparametric statistics carrier, including its objects, relations, parameters, conventions, evidence, boundary cases, and comparison targets, Every pair of observations is classified by whether the two variable orderings agree, disagree or tie, and a normalized signed difference measures ordinal association., and type the carrier, state every parameter and convention in the definition, test that the paired observations and sampling unit, ordinal variables and ordering, missingness, concordant discordant and tied-pair definitions, tau variant and denominator, estimate range, null hypothesis, variance or exact distribution and uncertainty are explicit, compare the nearest accepted identity, and report counterexamples, uncertainty, and limiting cases.

Relationships to Other Abstractions¶

Current abstraction Kendall rank correlation coefficient Domain-specific

Parents (1) — outgoing typed relations

Kendall rank correlation coefficient is a kind of Correlation Prime

It remains its own entry because its identity is fixed by the paired observations and sampling unit, ordinal variables and ordering, missingness, concordant discordant and tied-pair definitions, tau variant and denominator, estimate range, null hypothesis, variance or exact distribution and uncertainty.

Hierarchy path (1) — routes to 1 parentless root

Kendall rank correlation coefficient → is a kind of Correlation

Neighborhood in Abstraction Space¶

Kendall rank correlation coefficient sits in a crowded region of the domain-specific corpus (10^th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Descriptive Statistics & Correlation Measures (40 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25