JOHN TUKEY'S CONTRIBUTIONS TO MULTIPLE COMPARISONS

Abstract

This paper provides an historical overview of the philosophical, theoretical, and practical contributions made by John Tukey to the field of simultaneous inference. The Problem of Multiple Comparisons, released in 1953, provided not only the first comprehensive account of the field but also set much of the research agenda for the next 35 years. During the last decade of his life, Tukey devoted substantial attention to this area, experimenting with different graphical representations of multiple comparison procedures and exploring the implications of the false discovery rate (FDR) approach to controlling family-wise error rates. In a number of publications, Tukey continued to grapple with the fundamental issues of the field and to identify critical problems to be addressed.