Measuring Employment Discrimination through Controlled Experiments

Abstract

Race/ethnic discrimination in hiring can be measured under controlled conditions using matched pairs of minority and nonminority research assistants posing as applicants for the same job. In 149 in-person job applications in the Washington, D.C., labor market, African American applicants were treated less favorably than equally qualified nonminorities more than one-fifth of the time. Employer behavior during these interactions suggest that, within continued public and private efforts against discrimination, particular attention should be accorded to the cognitive underpinnings of bias.