Format

Send to

Choose Destination
See comment in PubMed Commons below
Biometrics. 1988 Sep;44(3):847-60.

Estimating the power of the two-sample Wilcoxon test for location shift.

Author information

1
Department of Mathematical Sciences, Montana State University, Bozeman 59717.

Abstract

Traditional methods for calculating the power of a statistical test for location shift require knowledge of the shape of the underlying probability distribution. The distribution shape, however, may be unknown. This paper describes a bootstrap method for using observed data (or pilot data) to approximate the power. No assumptions need be made about the shape of the underlying continuous probability distribution. Simulation evidence shows that, when applied to the Wilcoxon two-sample test for location shift, the suggested method is reliable. The evidence also shows that it is more accurate than a benchmark traditional approach. The bootstrap method is applied to a real-data example. The analysis demonstrates how the method can be used to determine sample sizes and how to choose the more powerful of two alternative tests for location shift.

PMID:
3203133
[Indexed for MEDLINE]
PubMed Commons home

PubMed Commons

0 comments
How to join PubMed Commons

    Supplemental Content

    Loading ...
    Support Center