Non- Parametric Statistics: A Set of Statistical Techniques to Compare Two or More Independent Populations

Abstract

Nonparametric methods are a powerful research tool used by investigators in practically every field of human activity. Nonparametric methods are useful alternatives to the parametric methods (when parametric methods are not available) and their use and application is made much easier when statistical tools, like MINITAB, are used to solve problems completely or partially. The techniques of we discuss in nonparametric statistics fall in the following 5 categories:

I) Tests for Randomness
II) Chi-Square Tests
III) Tests for Matched Pairs
IV) Tests to Compare 2 or More Independent Populations
V) Spearman Rank Correlation Test

MINITAB examples are given for: Tests of Randomness, Tests for Matched Pairs (Wilcoxon Sign Rank Test, Friedman Test) and Tests to Compare 2 or More Independent Populations (Mann-Whitney Test, Kruskal-Wallis H Test).