NCJ Number
90815
Date Published
1983
Length
279 pages
Annotation
Comparative case studies of Jacksonville, Fla., and San Diego analyze the amount of attrition for frequent, serious crimes (robbery, burglary, and felony assault); examine important factors that account for case attrition; and attempt to determine the effects of high attrition rates on the criminal justice system and whether strategies can be developed to decrease jurisdictions' attrition rates.
Abstract
The authors review problems of accurately measuring case attrition and defining arrests. Findings show how great differences exist between Jacksonville and San Diego in the way that arrests are made, that evidence such as identification and confessions is obtained, and the impact that different kinds of evidence have on convictions. Victim-witness problems are more frequent in San Diego but more damaging in Jacksonville, suggesting that police and prosecutorial agencies should pay greater attention to how they collect and use evidence. The single most important factor in determining a jurisdiction's attrition rate (high or low) is the arrest policy followed in that jurisdiction, while the single most important factor in determining a high or low conviction rate for a prosecutor's office is the office's policy regarding screening and filing. Among implications for police and prosecutors are that case salvageability depends upon earlier police investigation, greater investigative effort to solidify cases, and more risk-taking by prosecutors. Furthermore, police administrators need to spend more time on the problem of convictions, since convictions depend upon evidence produced by police; and prosecutors need to pay more attention to cases at an early stage. The empirical analysis is based on a review of prior research, letter and telephone contacts with more than 100 jurisdictions, brief visits to 10 sites, detailed observations in 4 locations, and extensive analysis of case records in Jacksonville and San Diego. Figures, numerous data tables, footnotes, and about 35 references are included. Sample cases and regression analysis methodology are appended.