Second Order Measures of Association — a Markovian Approach

Abstract

A technique previously developed for interpreting measures of association for fourfold tables based on markovian assumptions is here extended to derive measures of second order association. Transition rates between the states repre sented by 'the dichotomous cube" are decomposed in different ways to arrive at an additive, a multiplicative, and an exponential measure. The relative merits of these measures and their relation to traditional coefficients are discussed