확률화 블록 실험계획 모형에서 검정 통계량들의 검정력 분석

The Analysis of power of the Test Statistics for the Randomized Block Design

The purpose of this study is investigate the differences among parametric and nonparametric test statistics for the tree alternative hypothesis in the randomized block design. As the results, it was found that there was no large differences among parametric and nonparametric test statistics in power when the block sizes were larger, and Hollander's statistic had better power than other nonparametric test statistics. It is recommended that Hollander's test statistic is more useful method when we have no information about the distribution of population.