A document published by OECD in 1988 (IRRD 805303) defines an in-depth study as one which requires the collection of specific data in addition to data in the standard traffic accident record. These studies can be classified into statistically representative approaches and into clinical type studies. Care must be taken that collection and analysis procedures are not overstandardised. An in-depth accident study was carried out in France from 1980 to 1987. Data was collected on site as soon as possible, and analysis was based on a kinematic reconstruction of accident sequence. As a result, a method was developed for reconstructing and analysing accidents, and a databank was established. The sequential functioning model was developed by integrating cognitive psychology and artificial intelligence. Interviews with those involved are an indispensable part of data collection. The criticisms made regarding interviews concern: 1) account of somebody recently involved in an accident cannot be relied upon; 2) the nature of what can or cannot be memorised during such an event; and 3) the limitation of verbal material to gain access not only to the results but to the mental activity itself. Experience shows that onsite data collection is more reliable than deferred collection. Man is a being who rationalises, and does this very rapidly. For the paper by H.E. Pettersson see C 6520.
Abstract