Nonparametric Statistical Methods Using R

Preface

Nonparametric statistical methods for simple one- and two-sample problems have been used for many years; see, for instance, Wilcoxon (1945). In addition to being robust, when first developed, these methods were quick to compute by hand compared to traditional procedures. It came as a pleasant surprise in the early 1960s, that these methods were also highly efficient relative to the traditional t-tests; see Hodges and Lehmann (1963).

Beginning in the 1970s, a complete inference for general linear models developed, which generalizes these simple nonparametric methods. Hence, this linear model inference is referred to collectively as rank-based methods. This inference includes the fitting of general linear models, diagnostics to check the ...