What's happening on the network? : Railway accident modelling in a connected world

Introduction: Accident modelling, in an attempt to understand how the event happened, has traditionally being based on understanding accidents as a linear sequence of events and has not significantly evolved since its early inception in the safety sciences. Today however, modern software programs, particularly those being used for Social Network Analysis (SNA), offer the possibility to understand accidents and incidents as a complex network of relationships of contributing factors, perhaps for the first time. The analysis of railway incidents using this previously unexplored method has enabled accident investigations to move into a new era in the understanding of accident phenomenology in complex socio-technical systems. Aim: This research is interested in exploring how accident/incident data can be visually modelled using previously unexplored methods and what further understandings about the relationships between the contributing factors in railway incidents can be gleaned by the use of more modern methods of data (network) analysis.Method: Major railway incident reports, for a five year period (2006 – 2010), were analysed and data was collected on the contributing factors using the Contributing Factors Framework (CFF), a tool developed specifically for the rail industry in Australia. The contributing factors for various types of incidents were then modelled using Social Network Analysis methods.Results: The use of Social Network Analysis has enabled the relationship between of each of the contributing factors to be identified. The resulting models were analysed and the relationship between the main contributing factors for each type of railway incident was explored. By using this new methodology for accident analysis it is possible to obtain a clear visual picture of the inter-connectiveness of contributing factors and to further understand the relationship of constituent parts of complex systems. Discussion: The outcome of this research has been the development of a new method of accident modelling that moves beyond traditional linear models and better reflects modern day safety science thinking about complex socio-technical systems. It is hoped that this type of method can provide new and previously unknown knowledge about the interconnectedness of the contributing factors for different types of rail incidents and accidents in the quest to enhance accident prevention and learning.