Causal inference and racial bias in policing: New estimands and the importance of mobility data

2024-09-12

AI Summary

This paper addresses fundamental challenges in causal identification of racial bias in policing—namely, data limitations and untestable identification assumptions. We formally introduce the concept of “race-and-place policing,” construct an identifiable causal estimand, and develop a corresponding sensitivity analysis framework. We demonstrate that existing methods suffer from systematic bias when ignoring population dynamics and establish the necessity of mobility trajectory data for satisfying key causal identification conditions (e.g., ignorability). Empirical analysis using New York City data reveals statistically significant and robust racial disparities in policing, with estimates exhibiting strong robustness to violations of critical untestable assumptions. Incorporating mobility data substantially improves estimation precision. Our primary contributions include: (i) a novel theoretical model formalizing race-and-place interactions; (ii) a reconceptualization of identification conditions grounded in dynamic population exposure; and (iii) methodological extensions enabling more credible causal inference in observational policing studies.

Technology Category

• Reasoning under Uncertainty: Causality
• Machine Learning: Causal Learning
• Computer Vision: Bias, Fairness & Privacy

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Security and Privacy: Large-scale security measurements
• Social Networks and Social Media: Fairness and bias in social network and social media analysis

Abstract

Studying racial bias in policing is a critically important problem, but one that comes with a number of inherent difficulties due to the nature of the available data. In this manuscript we tackle multiple key issues in the causal analysis of racial bias in policing. First, we formalize race and place policing, the idea that individuals of one race are policed differently when they are in neighborhoods primarily made up of individuals of other races. We develop an estimand to study this question rigorously, show the assumptions necessary for causal identification, and develop sensitivity analyses to assess robustness to violations of key assumptions. Additionally, we investigate difficulties with existing estimands targeting racial bias in policing. We show for these estimands, and the estimands developed in this manuscript, that estimation can benefit from incorporating mobility data into analyses. We apply these ideas to a study in New York City, where we find a large amount of racial bias, as well as race and place policing, and that these findings are robust to large violations of untestable assumptions. We additionally show that mobility data can make substantial impacts on the resulting estimates, suggesting it should be used whenever possible in subsequent studies.

Problem

• Developing causal estimands for racial bias in policing
• Addressing race and place policing through mobility data
• Assessing robustness of racial bias estimates with sensitivity analyses

Innovation

• Developed new estimands for racial bias
• Incorporated mobility data into causal analyses
• Applied sensitivity analyses for assumption robustness