BOUNDS ON THE POWER OF LINEAR RANK-TESTS FOR SCALE-PARAMETERS

被引:6
|
作者
WASSERSTEIN, RL
BOYER, JE
机构
[1] KANSAS STATE UNIV AGR & APPL SCI,DEPT STAT,MANHATTAN,KS 66506
[2] KANSAS STATE UNIV AGR & APPL SCI,STAT LAB,MANHATTAN,KS 66506
来源
AMERICAN STATISTICIAN | 1991年 / 45卷 / 01期
关键词
ANSARI-BRADLEY TEST; FLIGNER-KILLEEN TEST; NONPARAMETRIC TESTS; SIEGEL-TUKEY TEST; SYMMETRICAL SCORE FUNCTIONS;
D O I
10.2307/2685232
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
We show that the power functions of a class of nonparametric tests for the equality of two scale parameters do not approach 1 as the ratio of the parameters approaches infinity. The class of tests, known as linear rank tests, is shown to have a fundamental flaw when applied to scale parameters, resulting in low power when the sample sizes are small.
引用
收藏
页码:10 / 13
页数:4
相关论文
共 50 条