Driving while black: do police pass the test?

Biased policing against racial and ethnic minorities is an important public policy issue. Theoretical analysis and empirical results on this issue has been plagued by an assortment of problems which confront research on the nature and significance of police discrimination against social groups. This paper presents and applies a nonparametric test that is robust to a host of methodological difficulties. We theoretically and empirically contrast our non-parametric test with other tests that are prominent in the literature. Utilizing data provided by the Florida Highway Patrol, our empirical results strongly reject the null hypothesis that FHP troopers of different races do not engage in racially biased searches of stopped drivers. More particularly, there is evidence of police bias against African American male and Latino drivers by all officers and no evidence of police bias against white male drivers by any group of officers.