Permutation Methods

The book will provide a comprehensive treatment of statistical inference using permutation techniques. Its purpose is to make available to practitioners a variety of useful and powerful data analytic tools that rely on very few distributional assumptions. Although many of these procedures have appeared in journal articles, they are not readily available to practitioners.

Most commonly-used parametric and permutation statistical tests, such as the matched-pairst test and analysis of variance, are based on non-metric squared distance functions that have very poor robustness characteristics. This second edition places increased emphasis on the use of alternative permutation statistical tests based on metric Euclidean distance functions that have excellent robustness characteristics. These alternative permutation techniques provide many powerful multivariate tests including multivariate multiple regression analyses. In addition to permutation techniques described in the first edition, this second edition also contains various new permutation statistical methods and studies that include resampling multiple contingency table analyses, analysis concerns involving log-linear models with small samples, an exact discrete analog of Fisher’s continuous method for combiningP-values that arise from small data sets, multiple dichotomous response analyses, problems regarding Fisher’s Z transformation for correlation analyses, and multivariate similarity comparisons between corresponding multiple categories of two samples.

Description of MRPP.- Additional MRPP Applications.- Description of MRBP.- Regression Analysis, Prediction, and Agreement.- Goodness-of-Fit Tests.- Contingency Tables.- Multisample Homogeneity Tests.- Selected Permutation Studies.