NCJ Number
77055
Date Published
1981
Length
105 pages
Annotation
This report describes a research study that developed evaluation methodologies to analyze crime commission rates and to further distinguish between low-rate and high-rate offenders.
Abstract
The basic model underlying the approach of the study was that there are 'K' types of crimes of interest, and each criminal offender commits each of the crimes at a specified rate (possibly zero) when free to do so. The study methods were viewed as helping to accomplish the following goals: (1) describe the distribution of the observed crime rates for the selected offenders who provided information about their actual crime commissions; (2) estimate the crime commission propensities of any one of these individuals, taking into account the group's overall distribution of commission rates as well as the individual's reported crime commissions and other characteristics; and (3) estimate the distribution of crime commission propensities for more general populations of offenders who differ from the selected offenders in known ways. The study describes obstacles to estimating the univariate distribution of commission rates for a particular crime type, including skewed distribution, too many zeros, and instrumentation error. It also examines the appropriateness of the mixed gamma-Poisson model for analyzing crime commissions and describes the adaptation of a procedure by Hudson and Tsui to estimate each individual's crime commission propensity from information about the number of crimes committed during the measurement period. Additional topic areas covered by the summary include multivariate modeling and extrapolations to more general populations of offenders. Footnotes, a figure showing the distributions of crime rate for robbery of persons, and 23 references are included. For an executive summary of this report, see NCJ 77054.