A distribution-free rank-like test for scale with unequal population locations
作者:
R. Clifford Blair,
G. L. Thompson,
期刊:
Communications in Statistics - Simulation and Computation
(Taylor Available online 1992)
卷期:
Volume 21,
issue 2
页码: 353-371
ISSN:0361-0918
年代: 1992
DOI:10.1080/03610919208813023
出版商: Marcel Dekker, Inc.
数据来源: Taylor
摘要:
The properties of a distribution-free rank-like test proposed by Moses (1963) for the twosample scale problem is studied and a modification of the test using Savage scores is proposed. It is shown that this rank-like test is superior to commonly used rank tests for scale in that it:(1) does not require the estimation of any location or centrality parameter, (2) does not require equal or known location parameters, (3) is robust for skewed data, (4) is resolving and (5) has some significant power advantages. The test is shown to be asymptotically normal, and asymptotic relative efficiencies are calculated. Power properties, studied via simulation, indicate that the test is especially well suited for testing for equality of scale when the data are sampled from skewed populations with unequal medians. Extensions to the J-sample problem are indicated.
点击下载:
PDF (550KB)
返 回