비모수검정의 비교연구 및 특정 분포에 대한 검정력 분석

Abstract

The Mann-Whitney test and the median test are two tests that can be used to test for a difference in location parameters. This paper compared powers of the two tests under a variety of population distributions through a simulation study. Both tests require that two underlying populations have the same variance, but this assumption was relaxed in some of the comparisons. In every case, equal sizes of 10 and 20 were used.

The Mann-Whitney test was found to be more powerful than the median test. It had the largest power when the underlying distributions were log-normal, beta, gamma, and chi-square for both sample sizes. The powers of the two tests were about the same for sample sizes of 10 and 20 from the mixture population 80% N(0,1) and 20% N(2,25) and for sample sizes of 10 from the mixture population 95% N(0,1) and 5% N(30,91). When the equal variance assumption was relaxed, the median test was found be more conservative than the Mann-Whitney test.