rankbased

Rankbased, or rank-based methods, refer to statistical techniques that base their computations on the ranks of data rather than the raw data values. They are a subset of nonparametric statistics and are particularly useful when data do not meet the assumptions of parametric tests, such as normality or homoscedasticity. By converting measurements to ranks, the influence of outliers is reduced and ordinal information is preserved.

Common rank-based tests for hypothesis testing include the Wilcoxon signed-rank test for paired samples, the Wilcoxon rank-sum test (also known as the Mann-Whitney U test) for two independent samples, the Kruskal-Wallis test for comparing more than two groups, and the Friedman test for blocked designs. Rank-based correlations include Spearman's rho and Kendall's tau, which assess monotonic relationships using ranks rather than raw values. In practice, rank-based methods are widely implemented in statistical software and can be applied to small samples or skewed data.

Advantages of rankbased methods include fewer distributional assumptions and greater robustness to outliers. Limitations include typically lower statistical power compared with well-specified parametric tests when their assumptions hold, and the interpretation of results in terms of ranks rather than raw measurements.

See also: nonparametric statistics, rank correlation, rank-based regression, nonparametric tests.