(1999) believes that many prosecutors have been slow to learn and apply modern principles of management and measures of accountability that seem warranted by their position of public power and responsibility. In the 1970s and 1980s, prosecutors implemented computerized case tracking and management systems in their offices, but few have used these capabilities to develop systematic information about case outcomes or management issues.
Part of the reason for this, as expressed by workshop participants, is a sense on the part of prosecutors that statistics cannot pick up what is important. For example, a great deal of work may go into a major financial fraud case that results in only a small number of indictments; yet the case may be very significant in terms of the further harm that would have been caused had the fraud gone unchecked. Also, as was once the case for police, the determination of what is actually being measured is an issue for prosecutors. Case outcomes, after all, are influenced not only by what the prosecutor does, but by the quality of the police investigation and the decisions of witnesses, juries, and judges.
As a result of these kinds of concerns, little systematic data on prosecution exists at the local or national level. For example, there is hardly any systematic information about prosecutor caseloads, number of crimes charged, cases bound over for trial, ratio of plea bargains to trials, or conviction rates. The National Judicial Reporting Program (NJRP), conducted by the Bureau of Justice Statistics, collects the most systematic data on conviction rates that we have today in the United States (Forst, 1999). The Bureau of Justice Statistics (BJS) has assembled conviction rates biennially from courts in about 300 counties around the country since 1986. However, these rates are not constructed from case tracking statistics, and so may distort the relationship between arrest and conviction for the six categories of crime covered, especially when the numbers of crimes are changing (Forst, 1999). BJS also conducts a National Survey of Prosecutors. Designed as a biennial series, the most recent available data were collected in 1996. The survey collects data on resources, policies, and practices of local prosecutors from a nationally representative sample of 308 chief litigating prosecutors in state court systems. It obtains basic information on staffing and operations and on current topics such as the use of innovative prosecution techniques, intermediate sanctions, juvenile cases transferred to criminal court, actions against prosecutors and other professional staff, and work-related assaults and threats (Bureau of Justice Statistics, 2000). In addition, BJS develops special reports on selected topics such as its re-