UMPU Test of Outliers in Random-effects Model of Two-way Nested Classification
Lei Shi · Chinese Journal of Applied Probability and Statisties · 1997
In this paper, we study outlier testing problem under the mean-shift model of main effects in balanced two-way nested classification. The likelihood ratio statistic is derived ,and its distribution under the null and alternative hypothesis are obtained respectiverly. A property of uniformly most powerful unbiasedness (UMPU) is proved. Finally some related topics are discussed.