Testing Statistical Hypotheses of Equivalence and Noninferiority (2nd ed.).

摘要

This second edition is a modest but welcome revision of the original. The book's title has been augmented to distinguish the case of "equivalence," normally associated with two-sided tests of an interval null, with "noninferiority," normally associated with related one-sided tests. Useful additions to this revised edition include a new closing chapter on tests for "relevant differences" and a switch from SAS and Fortran to R as the book's main computational engine. The book's outlook remains decidedly frequentist, and asymptotic arguments are used generously throughout. Bayesian approaches, quite natural for interval nulls that one hopes to accept rather than reject, are not covered. For these, the reader must turn to the books by Spiegelhalter, Abrams, and Myles (2004) and Berry et al. (2011).