Understanding Race Bias in the Decision to Shoot with an Integrated Model of Decision Making

Michigan State University, East Lansing MI

Investigators

Abstract

The shooting of unarmed citizens by police officers is a topic at the forefront of public awareness. The disproportionate rate at which unarmed Black citizens are shot is believed to reflect widespread racial bias on behalf of officers, which erodes public trust and reduces policing effectiveness. Calls to reduce bias with department-wide implicit bias training are common. There is evidence, however, that officers vary widely both in the rate of decision-making errors and in the degree of race bias in these decisions. Moreover, why race impacts decisions -- in terms of how it affects the underlying decision process -- varies across officers. Finally, the effectiveness of implicit bias training has been questioned recently, raising concerns that department-wide implicit bias training may not reduce unarmed shootings. These problems all reflect a larger theoretical problem: that the underlying cognitive and attentional components involved in an officer?s decision to shoot are not well understood. This project advances the basic science of decision making and does so in the context of a decision with great societal import. The overall goal of this project is to develop an Attention-integrated Model-based Shooting Simulator (AiMSS) to gain a deeper understanding of the mechanisms underlying the decision to shoot. The AiMSS combines computational models of decision making, visual psychophysics and eye-tracking methods, and an immersive decision simulator to map the processes police officers use to decide to shoot. The knowledge gained from AiMSS informs the development of training methods to minimize the impact of race on this decision and will serve as a generative tool to apply to other split-second decisions (e.g., medical decision making, security monitoring decisions). At the conceptual level, this research advances current computational modeling approaches by incorporating them with visual and social processing methods. This integration provides a more comprehensive understanding of this important decision that will ultimately aid in developing individualized interventions. Phase I will deploys a dynamic, immersive shooting simulator and mobile eye-tracking technology to collect behavioral and eye-gaze data during the decision to shoot. Using this rich data, the investigators develop computational models of decision making by integrating gaze behavior, pre-existing beliefs, and affective evaluations into an evidence accumulation process modulated by attention. An important strength of this approach is that both experts (sworn officers) and untrained participants will be tested. Phase II then uses results and methods from Phase I to compare different training interventions that seek to reduce the effect of race on the decision to shoot. The investigators develop the AiMSS to select individuals, identify the decision component responsible for biased responding, and deliver an intervention targeted to that specific component. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.