This report presents a summary of the development and testing of a method for multidisciplinary investigation at the scene of an accident. A literature study of more than seventy such investigations revealed that in many cases the outcome of these studies has fallen short of their objectives. A particular problem is that the results obtained are not amenable to generalisation. The need for a behaviourally based and transferable methodology which would emphasise the pre-crash phase was identified. The model developed described how the road-user functions in the traffic system, and provides a methodology for the collection of data and for analysing the course of events during the early stages of the accident. Using this model, 31 accident investigation case studies were undertaken. The methodology proved acceptable both in regard to the organisation of the at-scene investigations and the subsequent evaluation. One result of these case studies was that, unlike other such studies undertaken in the past, they generated recommendations directed at the road and traffic environment rather than the road user. In general, it is concluded, that while accident investigations have limitations, they represent a useful supplement to the conventional statistical analysis of accidents.
Abstract