Kendall's Tau Correlation Calculator

This Kendall's Tau calculator computes both Tau-a and Tau-b coefficients, showing the number of concordant and discordant pairs, and provides significance testing.

Further Information

Kendall's Tau is a non-parametric measure of the strength and direction of association between two ranked variables. Like Spearman's rho, it is useful when data violate the assumptions of Pearson's correlation.

Requirements

  • Two variables measured at least at the ordinal scale
  • Monotonic relationship between variables
  • No assumption of normality required
  • At least 3 pairs of observations

Tau-a vs Tau-b

Tau-a (simple formula):

Tau-a = (C - D) / [n(n-1)/2]

Tau-b (corrects for ties):

Tau-b = (C - D) / sqrt[(n0-n1)(n0-n2)]

Where C = concordant pairs, D = discordant pairs, n0 = total pairs, n1 = pairs tied on X, n2 = pairs tied on Y.

Concordant and Discordant Pairs

  • Concordant: Both variables change in the same direction
  • Discordant: Variables change in opposite directions
  • Tied: One or both variables have the same value

When to Use Kendall's Tau vs Spearman's Rho

Prefer Kendall's Tau when:

  • Sample size is small
  • There are many tied ranks
  • You want a more robust measure
  • Interpretation in terms of probability is preferred

Interpretation

  • Tau = +1: Perfect agreement (all pairs concordant)
  • Tau = 0: No association
  • Tau = -1: Perfect disagreement (all pairs discordant)

Note: Kendall's Tau values are typically lower than Spearman's rho for the same data.