Bayesian Networks for Probabilistic Inference and Decision Analysis in Forensic Science, 2nd Edition

Chapter 10Pre-assessment

10.1 Introduction

An evaluation process starts when the scientist first meets the case. It is at this stage that the scientist thinks about the questions that are to be addressed and the outcomes that may be expected. Together with the submitting agency or party, the scientist should attempt to frame propositions of interest that can help to think about the value of expected findings (Evett et al., 2000b). There is a wide tendency, however, to consider evaluation only as a final step in casework examination, notably at the time of preparation of the formal report. This is so even if an earlier interest in the process would enable the scientist to make better choices about the allocation of resources. One of the first approaches to strategic management in an operational forensic science unit was proposed by Cook et al. (1998b). It is based on a model embodying the likelihood ratio as a measure of probative value. The proposed model is intended to enhance the cost-effectiveness of casework activity from initial contact with the person or party requesting particular forensic services. The aim is to offer support in making better informed choices about which analyses to undertake.

In routine work, forensic scientists thus often require answers to questions such as ‘What are the magnitudes of the likelihood ratios that may be obtained?’, or ‘How probable is it that a likelihood ratio of a given magnitude will be obtained?’ A submitting party may wish to ...