Effect of racial misclassification in police data on estimates of racial disparities

Abstract

Research on race and policing increasingly draws upon data collected by police officers to estimate racial disparities in police contact. Many of these data sets, however, rely on officer perception of a stopped person's race, which may be inconsistent with how those individuals self-identify. Furthermore, researchers frequently benchmark contact data where race is perceived by police officers against census and survey data where race is self-identified. We argue that discordance between how individuals self-identify and how they are classified by officers can bias estimates of racial disparities. Using a unique data set, which allows us to compare officers’ racial classification of stopped persons with those same persons’ racial self-identification, we characterize rates of racial misclassification in administrative police records. We find evidence of racial misclassification in police records, especially among Hispanic and Asians/Pacific Islanders. We find that officer classification of Hispanics as (non-Hispanic) White is the most common form of racial misclassification in our sample and that its substantive consequences are significant. Specifically, we find that officer classification of Hispanics as White may lead analysts to incorrectly conclude that Hispanics are no more likely than Whites to be cited by police.