Suchergebnisse

Abstract The controversy of statistical evaluation of effectiveness of the present criminal justice system and of specific programs or changes implemented in the existing system has in part centered around the value of crime‐incidence data. The relationships between victimization and UCR or police‐blotter data pertinent to their use in evaluation are first described. An example is presented using an empirical‐stochastic model developed solely from UCR data to illustrate an approach to evaluation of system effectiveness. Here an ad hoc procedure is displayed which quantifies variation in and/or shifts in crime patterns after an intervention in the system.

AbstractThe power of the component randomization test for the paired sample location test is compared to the power of the parametric test for paired comparisons using the sum of differences distribution. This analysis is presented in the form of operating characteristic curves for normal, exponential, uniform and absolute lambda observations over a range of small sample sizes.

A dynamic descriptive model of the criminal justice system is implemented for the states of Missouri and Texas. The differences between sentencing strategies, prison population movements, and criminal behavior patterns are examined. Results for optimization of sentencing strategy and estimates for separate incapacitative and deterrent effects are provided from the discrete time model in each state, and comparisons are offered.

The impact on gun-related crime of legislative changes involving Massachusetts' Gun Control Law is evaluated. Viewing crime rates as realizations from some underlying stochastic processes, a procedure for measuring intervention effects on crime rates is presented. Time series on gun-related crimes for the city of Boston are analyzed with this procedure in evaluating the impact of the gun control law.

In this article, the characteristics of crime patterns in geographic areas over time are described by a network flow model. A method forforecastingfuture crime patterns with the network model is discussed along with procedures for evaluating future displacement by component randomization. An illustrative example of the model building and output analysis from actual arrest data is presented.