A Significantly Extreme Deviate in Data with a Non-Significant Heterogeneity Chi Square

Publisher Website
Google Scholar
Abstract
A test is given for detecting whether the extreme frequency of a set of binomial frequencies differs significantly from the others. Situations in which a phenomenon can only be recognized by the occurrence of such an extreme frequency are common, at least in the analysis of genetic data, and require a test of the sort given. It is pointed out that in spite of a non-significant overall heterogeneity X2, the extreme deviate may differ significantly from the remainder. Thus in general if departures of a specific type are to be expected then these must also be tested by an appropriate test of significance.

This publication has 1 reference indexed in Scilit: