Definition

1. Noun:
   • A nonparametric measure of partial correlation: Kendall partial rank correlation is a statistical method used to measure the strength and direction of the association between two variables, while controlling for the effect of one or more other variables. It is based on rank orders of the data rather than their raw values, making it a nonparametric technique.

Usage

• The Kendall partial rank correlation is used when the data do not meet the assumptions required for parametric partial correlation coefficients (like Pearson's), such as normality or a linear relationship.
• It is applicable for ordinal data or when outliers are a concern.
• The calculation involves assessing the concordance and discordance of ranks for the two primary variables, after adjusting for the influence of the controlled variable(s).

Examples

Advanced Usage

• Hypothesis Testing: The significance of a Kendall partial rank correlation coefficient can be tested, often using a standard normal (Z) approximation for larger sample sizes.
• Comparison with Other Measures: It is one of several partial rank correlation measures, alongside others like the Spearman partial rank correlation. The choice between them depends on the specific properties of the data and the type of rank correlation desired.

Variants and Related Words

• Kendall's tau (τ): The foundational rank correlation coefficient upon which the partial version is built. It measures the association between two variables without controlling for others.
• Partial correlation: A broader term for any statistical measure that quantifies the relationship between two variables while controlling for others. The Kendall partial rank correlation is a nonparametric type of partial correlation.
• Spearman's rank correlation coefficient: Another nonparametric measure of rank correlation, which has its own partial correlation variant.

Synonyms

• Kendall's partial tau
• Partial tau correlation

Notes

• This is a specialized term used primarily in the fields of statistics, data analysis, psychology, and social sciences.
• It is not typically used in everyday language or conversation.

Noun

1. a nonparametric measure of partial correlation