Decision Support and the Criminal Justice System

: an assessment of the potential for utilising technological decision support in undertaking the prosecutorial decision

Abstract

The continued development of technological tools for decision support offers potential for use within the criminal justice system. The Crown Prosecution Service (CPS) has been isolated as an organisation within which such decision support could be gainfully employed. The criticisms of the CPS since its inception in 1986 have been wide and varied, but many concentrate on the discretion of Crown prosecutors over the prosecutorial decision. This research represents a preliminary attempt to assess the feasibility of utilising decision support within the CPS to improve the quality and consistency of prosecutors’ decision-making.

Following an analysis of decision-making within the criminal justice system the problems and environment of the CPS are discussed in order to assess whether a decision support system could provide a source of improvement. The alternatives which have thus far been suggested are given and the proposed use of such decision support outlined as obtaining an independent assessment of the chances of conviction in a given case. An analysis is made of the various Artificial Intelligence (AI) traditions and techniques which could be adopted. It is submitted that three hold the most potential for the established objective: case-based reasoners, rule induction and artificial neural networks.

A practical experiment has been undertaken in an attempt to simulate one of these possibilities, the artificial neural network. A statistical analysis is described, together with the issues that were raised by examination of the data. Attempts to predict the outcome of cases utilising linear models are outlined and the results discussed. The methodology and results of the practical experiment are given and suggestions made for improvement and further research. In particular, the use of Sammon mapping is discussed with regard to the development of a hybrid system.

An interview study was also undertaken in order to assess the attitudes of those working within the criminal justice system towards the introduction of technological decision support in general and the simulated system in particular. The key issues arising from these interviews - the lack of technological knowledge within the criminal justice system as a whole, and the conflicting perspectives and objectives of different organisations and individuals - are discussed. Finally, conclusions are drawn and potential areas for further research highlighted.

Date of Award	1998
Original language	English
Awarding Institution	Aston University