Abstract In a nonparametric (distribution‐free) approach to the correlation between two sets of measurements made on the same individuals, each set may be ranked in order of magnitude. The standard (Pearson) product‐moment correlation coefficient is calculated on the two sets of ranks. Simple formulae are available, and adjustments are made for tied ranks. Test and estimation procedures are described. The relationship with the Kendall rank correlation coefficient is discussed.
Stored only when both papers are in the corpus — 591 citations exist upstream.