Segregation index: an orsametric for measuring discrimination against minorities

Orsametrics, the science of data and measurement appropriate to decision-making models, was proposed by H. M. Wagner in 1971. This paper proposes an orsametric measuring the distance between the current distribution of minorities in the organization and the worst distribution that minorities could have, given a specified proportion of minorities present in the organization. Measurement of this distance via the absolute value norm leads to a constrained model in which a convex, absolute-value functional must be maximized. This is accomplished by means of an algorithm designed to take advantage of the special structural properties of the model. An example is provided to demonstrate how the segregation index can be used to track discrimination levels within an organization through time. (Orsametrics: public sector applications-quantification of discriminatory hiring promotion practices; mathematical modelling-nonlinear, maximizing a convex function subject to linear constraints.