NCJ Number
217381
Journal
Journal of Quantitative Criminology Volume: 23 Issue: 1 Dated: March 2007 Pages: 1-22
Date Published
March 2007
Length
22 pages
Annotation
This paper presents a modified Poisson mixture model for estimating the factors involved in how chronic drug users get arrested, and this model is used to estimate arrest rates for 38 counties, using up to 16 quarters of data from the Arrestee Drug Abuse Monitoring survey.
Abstract
The proposed approach used data collected at booking facilities across 38 counties to estimate the rate at which chronic drug users in a county arrived at booking facilities, where they could be identified and counted. Estimating the rates at which chronic drug users arrive at these collection points allows researchers to include their drug use in estimates of drug use in the population as a whole. The Poisson mixture model (Cameroon and Trivedi, 1998; McCulloch and Searle, 2001)) is used to characterize the arrest process. Although this paper only discusses empirical estimates of the arrest process, the motivation for this research was to estimate the prevalence of chronic drug use in the community. A model of the arrest process is a preliminary step in weighting the sample of arrested drug users to represent drug use in the general population. This is necessary in producing a more reliable estimate of drug use. Conventional household surveys typically do not reach populations that are not accessible in household surveys. Chronic drug users must be surveyed where they congregate, i.e., in jails, treatment programs, and shelters. 3 figures and 33 references