Abstract
Abstract This article obtains the average effect on the estimator of Kendall's rank correlation coefficient when extreme pairs are removed from a sample. Formulas, which are functions of Sterling numbers of the second kind, are derived for this bias. The exact bias, and an approximation, are tabulated for n = 3(1) 19.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have