NCJ Number
79117
Date Published
1981
Length
31 pages
Annotation
The problem of bias in captive samples used for research on the relationship between drugs and crime is examined, with emphasis on a direct empirical comparison of data from a 1977-78 study done in Miami, Fla., and San Antonio, Tex.
Abstract
The possibility that captive samples are not biased must be rejected, because both the complexity of the relationships between crime and drug use and the nonuniformity of official handling of drugs/crime offenders means that criminally involved drug users are not all equally likely to end up in a captive sample. In addition, both measurement comparability problems and the need for multiple controls make it difficult to use a literature review to define the specific biases of captive drugs/crime samples. Thus, a direct comparison of drugs/crime data from captive and on-the-street samples was conducted using interview data from the 1977-78 study. A total of 942 heroin users were interviewed. Most respondents were active heroin users interviewed on the street, while the rest were users from the same communities who were interviewed in jails and treatment centers. The Miami sample included blacks and whites, while the San Antonio sample consisted of Mexican Americans. Twenty variables were chosen for use in comparing the active and captive samples. Although the active and captive samples displayed some strong similarities, they showed an even larger number of significant differences. It was concluded that findings from captive samples cannot be generalized to all criminally involved drug users; there is need for using multiple controls for active/captive comparisons. Findings also indicated that the past and current relationships between drug use and criminal behavior show differences in all the active/captive comparisons. Additional implications, tables, notes, and a list of nine references are provided.