Misapplications of the Wilcoxon-Mann-Whitney Test
Research Mentor(s)
Noguchi, Kimihiro
Description
Experimental designs using two or more treatments frequently arise in many fields of study, from medicine to psychology to policymaking. The use of rank-based nonparametric methods is recommended for analyzing effects in these designs as the observed data typically come from non-normal distributions with possibly several outliers. One of the most commonly used technique when conducting nonparametric analysis on two independent treatment groups is the Wilcoxon-Mann-Whitney (WMW) test, which is often thought of as a comparison between population medians of the two treatment groups. It can be shown that the WMW test actually fails as a comparison of medians and is better suited to be a comparison of stochastic superiority between these two groups.
Document Type
Event
Start Date
18-5-2020 12:00 AM
End Date
22-5-2020 12:00 AM
Department
Statistics, Mathematics
Genre/Form
student projects, posters
Subjects – Topical (LCSH)
Statistical hypothesis testing; Medical statistics
Type
Image
Rights
Copying of this document in whole or in part is allowable only for scholarly purposes. It is understood, however, that any copying or publication of this document for commercial purposes, or for financial gain, shall not be allowed without the author’s written permission.
Language
English
Format
application/pdf
Misapplications of the Wilcoxon-Mann-Whitney Test
Experimental designs using two or more treatments frequently arise in many fields of study, from medicine to psychology to policymaking. The use of rank-based nonparametric methods is recommended for analyzing effects in these designs as the observed data typically come from non-normal distributions with possibly several outliers. One of the most commonly used technique when conducting nonparametric analysis on two independent treatment groups is the Wilcoxon-Mann-Whitney (WMW) test, which is often thought of as a comparison between population medians of the two treatment groups. It can be shown that the WMW test actually fails as a comparison of medians and is better suited to be a comparison of stochastic superiority between these two groups.