Abstract
The purpose of the present study is to examine traffic stop data—in particular data on searches—to better understand racial profiling. We use a measure of capriciousness to understand whether the decision to search occurs at random. Using data from more than 40,000 traffic stops from Louisville, Kentucky, the present study finds that race is one of many factors that are used in the determination of a search. For the overall, White, and Black samples, we find that the decision to search occurs at random, suggesting that racial profiling is not occurring in these data.
Notes
It should be noted that this department no longer exists. On January 6, 2003, a merger of the Louisville and Jefferson County governments was conducted as a result of a popular vote. The new department, the Louisville Metro Police Department, now serves all of Jefferson County.
If we were to use ordinary least squares with these data, we would violate a major assumption that our dependent measure was captured at the interval or ratio level; thus, because the dependent measure is dichotomous, we used logistic regression (Menard, 1995).