Estimating Ethnoracial Impacts for Outcomes of Post-Stop Decisions

• In Event: Stop and Frisk: Some Outstanding Questions

Abstract

Police may frisk and/or search stopped pedestrians or drivers, and such post-stop actions may or may not surface contraband and/or weapons. Many scholars have heatedly debated appropriate and inappropriate models and assumptions for gauging ethnoracial differentials for frisk or more commonly search “hit” rates. The current paper suggests a simple empirical and conceptual point: simple hit rate analyses following the Knowles, Persico and Todd (KPT) model may be conflating three distinct but related dynamics: stop initiation, selecting for frisk or search, and determining the outcome of the selection decision. Conflating versus separating those dynamics may result in markedly different estimates of race and ethnicity impacts on hit rates. Also in question is how to conceptually organize the contributions of other features of context, stop, and civilian to each of these three dynamics. These points may be illustrated using data from a major city in the United States.

Author

• Ralph B Taylor, Temple University