Simultaneous Inference Using Finite Intersection Tests: A Better Mousetrap.

In the behavioral and social sciences researchers utilize classical simultaneous test procedures (STP) developed by Duncan (1947, 1955), Scheffe (1953), Tukey (1949, 1953), and Roy and Bose (1953), and others to analyze differences among means. In this article we bring to the attention of the applied researcher the obscure finite intersection test (FIT) procedure developed by P.K. Krishnaiah ( 1964, 1965a. 1965b. 1965c), compare the method with more familiar classical procedures, and illustrate how one may use the FIT procedure to analyze differences among all means for both univariate and multivariate experimental designs.