Methods of Measuring and Detecting Discrimination in Punishment

The law concerning racial discrimination in punishment has never been established clearly in legislation. Instead, through a series of court decisions, a constitutional basis for challenging alleged discriminatory punishment has evolved. In that evolution, the issue of substance has turned out to hinge on statistical “proof” of discrimination. Few, if any, challenges have succeeded. This is due, in part, to the proliferation of inappropriate statistical methods for measuring discrimination found in the social science literature dealing with punishment disparities. The problem is compounded further by the fact that a recent National Academy of Science report on sentencing disparities failed effectively and correctly to resolve the issue of how to detect or measure punishment discrimination,(Blumstejn. et. al.) This paper provides some simple illustrations to help clarify the role of statistical theory and analysis in the legal proof of racially discriminatory punishment. It is useful to begin by summarizing the chronology of the landmark decisions and past statistical analyses. By explicitly considering the legal context of several key punishment disparity cases, the statistician will gain a better understanding of the limitations of conventional methods of measuring racial discrimination in punishment