Reducing biases in the criminal legal system: A perspective from expected utility

Abstract

Objective Racial biases exist in almost every aspect of the criminal legal system, resulting in disparities across all stages of legal procedures-before, during, and after a legal procedure. Building on expected utility theory, we propose an expected utility framework to organize and quantify racial disparities in legal procedures. Hypotheses Corresponding to the parameteres involved in estimating expected utility, we hypothesized that racial biases would occur at different stages of legal procedures. Method Using police interrogation procedures as an example, we obtained estimates from previous literature and demonstrated that racial disparities exist at each stage of legal procedures. We then used these estimates to compute and visualize expected utilities, which quantify the average long-term outcomes of interrogations for minority versus nonminority suspects. Results Based on this hypothetical example, the expected utility analysis suggests that biases at various stages of interrogations could potentially lead to substantial disparities in legal outcomes between racial groups. In particular, the example shows that interrogations might yield notably worse outcomes for minority suspects than nonminority suspects because of cumulative biases that occur before, during, and after this legal procedure. Conclusions The proposed expected utility approach not only offers a valuable tool for accounting the joint impacts of multiple stages of legal procedures to quantify racial disp…