Police Forensic science performance indicators – a new approach to data validation

Abstract

DNA and fingerprint identifications continue to form an integral part of the detection of a wide range of crime types, especially volume crime such as burglary and auto crime. More than ten years ago, researchers first commented on the lack of emphasis on `outcome' (i.e. crime detection) related performance indicators for UK police forces. Since then much work has been carried out, mainly by the Association of Chief Police Officers of England & Wales and the Home Office, to produce a framework of forensic science performance indicators that reflect accurately the contribution made by forensic science to crime detection. In this paper, we consider the data currently being collected by five UK police forces that use popular proprietary computer based data collection systems. The accuracy of the data collection has been analysed using a neural network and has identified collection errors in all five forces. These errors are such that they could adversely affect the accuracy and interpretation of the national collection of forensic science data conducted by the Home Office. We propose using this neural network to check the accuracy of data collection and also to provide a `front end' collator for national forensic science data returns to the Home Office. Such an approach would improve the accuracy of data collection nationally and also provide some reassurance over the consistency of data recording by individual forces.