Race-Sensitive Choices by Police Officers in Traffic Stop Encounters

Abstract

This study introduces a statistical estimator that can be used to examine disproportionate traffic stop behavior of police officers. This estimator can be employed in concert with internal benchmark data and a tree diagram algorithm to identify and classify disproportionate behavior. These methodologies are multilevel and can be used (a) at the macrolevel to examine disproportionality of a police department as an organization and (b) at the microlevel to draw inferences about reasons for individual officers’ disproportionate behavior. These statistical routines were tested using data from a medium sized midwestern community. Results suggest that the models are effective in detecting disproportionality in both a police organization and an individual officers’ traffic stop activity. Moreover, the methods may serve as an initial step in pointing toward the sources of the officers’ behavior.