An example from the Banbury Enquiry of a three-dimensional contingency table for the chi-square test

Abstract

A recent book by David Newbold (1977) on the Banbury Enquiry into Ability Grouping contains an analysis of children's choice of friends. The variables concerned are sex, mixed ability/streamed classes, and closeness of ability of the friendship pairs as measured by a verbal reasoning test at the end of the first year of secondary schooling. The method of analysis was such that all three variables were dichotomous. This article re-analyses Newbold's data by using a little known development of the chi-square test and suggests that this technique deserves to be more widely known in educational research.