Development and validation of a theory of construction accident causation

Construction accidents may involve complex causal processes. Removing or mitigating the causes of construction accidents requires a comprehensive understanding of why accidents happen and how their causal processes work. In investigating these two different, but vital questions, the aim of this research was to develop and test a theory of construction accident causation.Through extensive literature research and discussions with safety experts in construction, the research, conducted in the Department of Civil & Construction Engineering, UMIST, United Kingdom, has produced The Constraint-Response Theory of Construction Accident Causation. It is a fundamental concept of representation of how and why construction accidents occur. The concept is that each project participant operate within variety of constraints, arising from features of the project environment or produced by the behaviour of other project participant, and their responses to the constraints can generate inappropriate situations or conditions, which directly increase the likelihood of accident events on construction sites. This concept is then presented into a constraint-response model, representing the way in which constraints experienced by project participants, and their responses, can introduce deficiencies of construction processes leading to accidents. The description of the model is subject to the project development organisation and its process, beginning with the constraint-response generation by clients, the client’s project team, designers, and contractors during project conception, design and throughout construction.In this thesis, the development process of the theory is reported. Philosophical issues in developing a causal theory of construction accidents and an overview of past attempts at development of accident theories are discussed. The principles of the Constraint- Response Theory, the causal model representing the theory, and the identification and classification of individual causal factors, including into proximal and distal classes, are described. The theory was tested by analysis of 1000 recorded construction accident cases provided by The UK Health & Safety Executive (HSE) from its FOCUS database. Since inadequate representation of the accident data, both quality and quantity of description of accident events, in this research, the theory is only partly tested. In this case, only features of the proximal factors in the theory was tested by statistical analysis of the text of the records of construction accidents. This showed that: a) 95% of individual proximal factors included in the theoretical model had been found in real construction accidents; b) many of the causal relationships postulated in the model were tested by statistical measures of association. Test of the distal factors is being continued through a HSE funded project to study 100 recent construction accident cases in collaboration with Loughborough University.It was found that these currently available data are, in some respects, qualitatively and quantitatively inadequate and will need to be supplemented in the future, by extended accident investigations. A more structured and detailed investigation process would promote a clearer understanding of the relative importance of all causal factors and their relationships. This is essential if the full accident causal process is to be properly understood and evaluated. Through all this research effort, it is expected that the theory will improve understanding of the accident causation process, assist in the structured investigation of accidents, and offer guidance on effective accident prevention measures.