Abstract
Using 1989 data on illegal drug possession and trafficking arrests and socioeconomic data from the 1990 census, this article presents a series of models of drug arrest rates for a sample of U.S. cities. Comparing predictions derived from social disorganization and conflict theories, the models demonstrate that the racial composition of these cities, even while controlling for economic deprivation variables, is the strongest predictor of drug arrest rates. In addition, and also consistent with conflict theory, a measure of police force strength exerts a significant positive effect on trafficking arrest rates.