NCJ Number
92960
Date Published
Unknown
Length
206 pages
Annotation
This report describes the statistical methods and results of a crime and arrest forecasting study for Alaska and Oregon which was based on several years of statistics from 16 Alaska and 39 Oregon Uniform Crime Reporting areas. This research also addressed the feasibility of forecasting in a routine agency context.
Abstract
The first section discusses the state-of-the-art of crime forecasting and identifies areas where it may be used as a planning tool. An analysis of the literature shows that crime forecasting has gained in popularity, but a national survey of criminal justice agencies indicates that its practical application is not widespread. Conceptual considerations for univariate based and causal based forecasting are discussed, and a social policy framework is offered that sets forth requirements for a multimethod approach to forecasting crime and arrests in Alaska and Oregon. The next section describes the research setting, measures, data collection procedures, and problems encountered. A comparison of three distinct forecasting methods -- ARIMA or Box-Jenkins time series analysis, econometric or structural equation modeling, and panel or pooled time series analysis -- concludes that the ARIMA method is the least costly. Finally, the report presents a step-by-step discussion of the forecasting methods and the results. In summary, the study concluded that crime rates in Oregon and Alaska will increase through 1987, but that in both States the bulk of this increase will be in property crime and in Alaska the increase will be realized almost entirely in Anchorage. The analyses also uncovered lead indicators of crime and arrest rates, such as the criminal filing rate and imprisonment rates, which in theory the policymaker can manipulate. The report includes maps, graphs, tables, approximately 80 references, and materials from the national survey on criminal justice.