A Simple Simultaneous Test Procedure for Quasi‐Independence in Contingency Tables

This paper presents a simple procedure for testing simultaneously a series of hypotheses pertaining to whether or not the row and column classifications of an R × C cross‐classification table are “quasi‐independent”. Applied to a given table, this procedure will limit the chances of rejecting any hypotheses about quasi‐independence that are actually true, even when the series of hypotheses that are investigated have been suggested to the research worker by earlier scrutiny of the same data, and also when these hypotheses have been specified a priori. to illustrate this procedure, we reanalyse data on British social mobility which describe the cross‐classification of each subject's occupational status category and his father's occupational status category.