Arrest Probabilities for Marijuana Users as Indicators of Selective Law Enforcement

Abstract

This paper estimates arrest probabilities for marijuana users by combining detailed arrest statistic for three metropolitan areas (Cook County, Ill,; Douglas County (Omaha), Nebr., and the Washington, D.C., metropolitan area) with marijuana user rate estimates from questionnaire surveys in those three jurisdictions. The analysis involves a comparison of arrest probabilities for several sociodemographic categories of marijuana users in these jurisdictions. The baseline model for these comparisons is the equiprobability hypothesis that seems implicit in much criminal justice research: relative to violation rates, the likelihood of sanctions ought to be equal. The data show that arrest probabilities for marijuana users vary by gender and occupation in all three jurisdictions; variations by age and race are more jurisdictionally specific. These arrest probabilities are then examined by introducing additonal data pertaining to the circumstances of detection and arrest. The observed differences in arrest probabilities are interpreted to be compatible with two alternative explanations which have rather different implications for understanding selective law enforcement.