University of Victoria (Canada)
A review of nonparametric statistical tests for randomized block designs
Abstract
dc:description.abstractIn this thesis we examine some commonly used nonparametric rank tests for the situation where experimental data is given in the form of a two-way layout from k treatments and n blocks. The null hypothesis is that there are no treatment effects. We consider statistical test procedures for general alternatives and for ordered alternatives; and multiple comparison procedures for all treatment and for treatment versus control comparisons. Large sample approximations and illustrative practical examples are given. We derive moments of the relevant statistics, justify the large sample approximations, consider derivation of tabular values, and review the consistency and asymptotic relative efficiency properties of the tests. Several selected test procedures for more general randomized block designs are discussed. These include tests for general alternatives in balanced incomplete block designs and in designs with an equal number (~1) of treatment observations per cell; a test for ordered alternatives using cell location parameters; and a test for ordered alternatives based on a weighted sum of Jonckheere statistics.
Degree
thesis:*- Level thesis:degree_level
- Master of Science M.Sc.
- Department thesis:degree_department
- Department of Mathematics
- Year dc:date.issued
- 1980
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Geary, David Norman
Rights
dc:rights- Statement dc:rights
-
- Available to the World Wide Web
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/1828/17871