A Study On Detecting Outliers In Two-Way Tables

이원배치법(二元配置法)에서의 이상치(異常置) 발견방법(發見方法)에 대한 연구(硏究)

Basic problems in the study of detecting outliers from data of experimental designs are that they are difficult to detect and their presence influences the analysis of variance of the data set. This article is concerned with mainly detecting outliers in two-way tables with no replications. Various methods are reviewed and their relations to the Andrews-Pregibon's Statistic and Cook's Statistic are derived.