The Color of Confinement: Racial Bias and Jail Populations Across America

Abstract

This study builds on the body of research examining whether racial disparities in criminal justice can be attributed to bias. The purpose of the current study was to examine whether there is a relationship between aggregate levels of bias and race-specific incarceration rates in U.S. counties. With data from the Vera Institute of Justice, the U.S. Census Bureau, and the Harvard Project Implicit, this study uses county-level estimates of implicit and explicit biases via Multilevel Regression with Poststratification to assess the relationship between those two types of biases and Black and White prisoners in 2,825 county jails across the U.S. using negative binomial regression. Results indicate that pro-White/anti-Black explicit and implicit bias are associated with a higher population-adjusted number of Black prisoners, and fewer White prisoners, even after controlling for socioeconomic covariates and arrest rates. This research provides compelling evidence that racial bias may contribute directly to racial inequity in jail populations and that bias can be understood as a collective phenomenon impacting social systems.