Some Extensions

Abstract

We deal with certain extensions of the theory of nonparametric tests, which have not been covered in previous chapters. In particular, we consider linear rank tests for one-sample problems. Under a location parameter model, two pseudo-groups can be constituted by the observables which are smaller or larger than the parameter value under the null hypothesis, respectively. Resulting rank tests, like the sign test and Wilcoxon’s signed rank test, are discussed. Furthermore, we address the problem of tied observations in rank-based statistical inference. It is recommended to assign midranks in the presence of ties.