Analysing road traffic crash causation from a systems-based perspective: applying the Driver Reliability and Error Analysis Method

This thesis comprises eight peer reviewed publications which were published between 2008 and 2020. Six of these present analyses using the Driver Reliability and Error Analysis Method (DREAM) crash causation method. The remaining publications employ two other systems-based crash causation methods to explore the potential benefits of non-DREAM alternatives. These publications represent a number of contributions to the application and further development of methods over the course of the 12-year research programme, which have helped to build a more systemic understanding of road crash causality. This thesis demonstrates the new contributions to knowledge made by the application of DREAM, and subsequent refinement of the analysis procedures, to large datasets and less ideal data sources. In the first part of the thesis the application of the DREAM method to a large European road crash causation database is described and evaluated. The publications presented establish procedures for applying the DREAM method and demonstrate its effectiveness in describing crash causation. The second part of the thesis documents the analysis of data from UK fatal collision investigation reports using two alternative systems-based methods (a project specific case-study based method and AcciMap) before exploring the potential role of DREAM to analyse such data. This was the first time DREAM had been applied to data generated from UK fatal collision investigation reports and represents an expansion of the DREAM method. This thesis demonstrates that DREAM is an effective methodology in identifying the contributory factors of crashes within large European datasets and can be used to address a variety of different research questions. With its use within the European Commission supported projects SafetyNet, DaCoTA and SaferWheels, DREAM is the most widely applied systems-based causation analysis methodology in road safety. This thesis demonstrates that DREAM can be used as a valuable tool to allow systems methodologies to be applied within a road safety context by producing systematic crash causation analyses. The individual publications included in this thesis also provide examples of the new knowledge that can be generated by conducting analyses, for example, the role of inattention and distraction in crashes and the main causes of Powered Two-Wheeler collisions that occur at junctions. In the future the DREAM taxonomy could be combined with other established systems based accident causation methodologies such as AcciMap to provide broader knowledge on road crash causation.