Tests for bivariate symmetry against ordered alternatives in square contingency tables

Abstract

Let X and Y denote two ordinal response variables, each having I levels. When subjects are classified on both variables, there are I-2 possible combinations of classifications. Let p(ij) = Pr(X = i, Y = j). This paper introduces a family of tests based on phi-divergence measures for testing H-0: p(ij) = p(ji) against H-1: p(ij) greater than or equal to p(ji) (i greater than or equal to j); and for testing H-1 against H-2: p(ij) unrestricted. A simulation study assesses some of the family of tests introduced in this paper in comparison to the likelihood ratio test.