THE UNBIASEDNESS OF HODGES AND LEHMANN TEST FOR A COMPOSITE HYPOTHESIS OF THE MEAN OF A NORMAL-DISTRIBUTION WITH UNKNOWN-VARIANCE

被引：4

作者：

FRICK, H

机构：

[1] Institut für Datenanalyse und Versuchsplanung, Wessobrunnerstr. 6 D-8035, Gauting

来源：

COMMUNICATIONS IN STATISTICS-THEORY AND METHODS | 1990年 / 19卷 / 08期

关键词：

composite hypotheses; similar tests; t-test; unbiasedness;

D O I：

10.1080/03610929008830349

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 ; 070103 ; 0714 ;

摘要：

In 1954 Hodges and Lehmann considered the following problem: given is an i.i. normally distributed random sample with variance unknown. Under the null-hypothe-sis the mean is contained in a prescribed interval. Hodges and Lehmann constructed a test similar on the boundary, i.e. the level does not depend on the unknown variance when the mean is an endpoint of the interval. This test is superior in power to the usual auxiliary procedure applied to this problem. Numerical calculations by Hodges and Lehmann indicated that the test is unbiased, however an analytical proof could not be given. In a recent paper the author proved un-. © 1990, Taylor & Francis Group, LLC. All rights reserved.

引用

页码：2799 / 2814

页数：16