Analyzing personnel selection decisions in employment discrimination litigation settings

Abstract

In many employment discrimination cases, the employment outcome that is the subject of the complaint involves whether employees from a protected group receive the benefit of a particular decision at the same rate as the members of a presumably preferred group. For example, applicants can be hired or not hired, promoted or not promoted, terminated or not terminated, and so on, and the rates of attaining these outcomes (whether positive or negative) may differ across groups. These decisions are evaluated in a substantial portion of all employment discrimination claims (Zink and Gutman, 2005).This chapter familiarizes readers with how to analyze personnel selection decisions in employment discrimination litigation. Toward that end, we first outline some of the basic legal principles that serve as the basis for analyses related to claims of discriminatory employment practices. Second, we describe how to conduct a scientific investigation of the merits of such claims. This section focuses on analyses of disparity using traditional applicant flow data, and we present both statistical significance tests such as Fisher's exact test and practical significance tests including the four-fifths rule. To better illustrate the use and interpretation of these analytical strategies, we present results with simulated data that include a dichotomous 67