Permutation Tests : A Practical Guide to Resampling Methods for Testing Hypotheses

This book presents a readable introduction to the use of permutation methods for testing statistical hypotheses. The chief value in using such methods is that they may be applied in a wide variety of settings, to continuous, discrete, and categorical data, and to values from almost all probability distributions. In addition, permutation tests are generally robust in circumstances where there are problems with missing data. The emphasis of the book is on the practical use of such tests, and throughout examples are given to illustrate the techniques discussed. These examples are drawn from a wide range of settings including archaeology, biology, climatology, education, and the social sciences. An extensive bibliography is also included which provides references to over 500 articles related to the subject of permutation tests. As a result, many researchers will find this an invaluable reference work, as well as a handy desktop companion.

This text will equip both practitioners and theorists with the necessary background in testing hypothesis and decision theory to enable innumerable practical applications of statistics. Its intuitive and informal style makes it suitable as a text for both students and researchers. It can serve as the basis a one- or two-semester graduate course as well as a standard handbook of statistical procedures for the practitioners’ desk.

Parametric, permutation, and bootstrap procedures for testing hypotheses are developed side by side. The emphasis on distribution-free permutation procedures will enable workers in applied fields to use the most powerful statistic for their applications and satisfy regulatory agency demands for methods that yield exact significance levels, not approximations. Algebra and an understanding of discrete probability will take the reader through all but the appendix, which utilizes probability measures in its proofs.

The revised and expanded text of the 3rd edition includes many more real-world illustrations from biology, business, clinical trials, economics, geology, law, medicine, social science and engineering along with twice the number of exercises. Real-world problems of missing and censored data, multiple comparisons, nonresponders, after-the-fact covariates, and outliers are dealt with at length. New sections are added on sequential analysis and multivariate analysis plus a chapter on the exact analysis of multi-factor designs based on the recently developed theory of synchronous permutations.

 The book's main features include:

Detailed consideration of one-, two-, and k-sample tests, contingency tables, clinical trials, cluster analysis, multiple comparisons, multivariate analysis, and repeated measures

Numerous practical applications in archeology, biology, business, climatology, clinical trials, economics, education, engineering, geology, law, medicine, and the social sciences

Valuable techniques for reducing computation time

Practical advice on experimental design

Sections on sequential analysis

Comparisons among competing bootstrap, parametric, and permutation techniques.

 From a review of the first edition:

"Permutation Tests is a welcome addition to the literature on this subject and will prove a valuable guide for practitioners . . . This book has already become an important addition to my reference library. Those interested in permutation tests and its applications will enjoy reading it." (Journal of the American Statistical Association)

From a review of the second edition:

"Permutation Tests is superb as a resource for practitioners. The text covers a broad range of topics, and has myriad pointers to topics not directly addressed. . . the book gives guidance and inspiration to encourage developing one’s own perfectly tailored statistics...The writing is fun to read." (John I. Marden)