Tests of independence for multidimensional contingency tables based on (h − ,φ − − ) -divergence measures

Abstract

A divergence statistic is used to quantify the difference between a joint probability and the product of its marginal probabilities. An application of this to categorical data analysis is given. The asymptotic distribution of this statistic, and the Pitman and Bahadur efficiencies of tests of independence based on this statistic are obtained.