Unpacking gender and racial/ethnic biases in the federal sentencing of drug offenders: A causal mediation approach

Recently scholars have noted a limitation with current sentencing research is that the distinct contributions of prosecutorial and judicial discretion to sentence length have not been quantified. The current study seeks to calculate the proportion of sentence length disparities across gender and race/ethnicity that can be attributed to prosecutorial-initiated departures and sentencing decisions. Using data on federal narcotics offenders, the current study employs recently developed causal mediation analytic methods to decompose the direct and indirect influences of gender and race/ethnicity on sentence length. Results indicate the decisions of judges and prosecutors contribute both to gender and racial/ethnic disparity in outcomes; however, judicial decisions are mainly responsible for gender disparities in sentence length whereas prosecutorial and judicial discretion contribute more equally to racial/ethnic disparities. Reducing extra-legal disparity requires evaluating decisions of actors at different decision points. Abatement of gender disparities necessitates a primary focus on judges, but some attention to prosecutorial decisions is still warranted. Addressing race/ethnic differences in sentence length requires focus on both the decisions prosecutors make in granting departures and the ultimate sentence that judges decide to levy against the defendant. •Causal mediation analysis is used to decompose sources of sentence length disparities.•Judges contribute predominately to gender sentence length disparities.•Judges and prosecutors contribute equally to racial/ethnic sentence length disparities.