Permutation Tests for Nonparametric Statistics Using R

The R-package jmuOutlier, which contains functions for performing nonparametric statistical analyses, is introduced. This R-package is quite simplistic, and is intended to accompany undergraduate textbooks on nonparametric statistics. The main objectives of this R-package include the following: using R for running a permutation test on a mean, median, or any other statistic; running a permutation F-test on the difference between location parameters, without using an asymptotic approximation to the F-distribution; running a permutation test on Pearson and Spearman correlation, without using an asymptotic approximation to the normal or t-distribution; running the permutation Siegel-Tukey test in both the forward and reverse directions and the ratio mean deviance test on the difference between scale parameters. Additionally, this R-package provides exact power calculations using the binomial test, provides exact con dence intervals on percentiles (based on the binomial test), plots two empirical cumulative distribution functions on the same graph, and produces van derWaerden and exponential scores. The aim is to perform these tests based on either all permutations or a large number of simulated permutations, to obtain p-values without using asymptotic approximations. The greatest contribution of this R-package is simplicity, in that an elementary understanding of R is sucient when using this package. We conclude that students can both learn nonparametric statistics and perform analyses using this R-package, with just a basic understanding of R.