Date of Award:

5-1988

Document Type:

Thesis

Degree Name:

Master of Science (MS)

Department:

Mathematics and Statistics

Department name when degree awarded

Applied Statistics

Committee

David L. Turner

Abstract

This paper discusses results of a computer simulation to investigate several different tests when sampling several distributions. The hypothesis H0: μ=0 was tested against H0: μ≠0, using the usual t-test, trimmed t-test, the Jackkinfe, the Boostrap and signed-rank test. The p-values and empirical power show that the Bootstrap is as good as the t-test. The Jackknife procedure is too liberal, always obtaining small p-values. The signed-rank is a fairly good test if the data follows the Cauchy Distribution.

Checksum

cd47897862f039bb5b3356b252b7b7a0

Share

COinS