Is the longbow better than the crossbow? Emerging issues from mobilising a longitudinal study on a megaproject

Longitudinal studies of occupational safety and health (OSH) outcomes in construction projects are rarely conducted, due to the financial, practical and ethical difficulties of studying people, projects, and organisations over extended periods of time. Traditionally, OSH research in the construction industry is cross-sectional – where a ‘snapshot’ is taken, often with a retrospective view. The focus of this paper is the mobilisation of a longitudinal research study investigating OSH policy in an eight-year infrastructure megaproject in the UK. The research examines implementation of the project’s “transformational” OSH strategy, in order to develop new understandings of the effectiveness of OSH interventions. The research design uses a “strategy as practice” lens and traces the various strands of OSH policy, from development to their adoption as practice. The research context is complex, due to the complicated contractual arrangements. The research design incorporates a rarely used “tracer” methodology. During the mobilisation phase of the research project, several challenges were identified, including interpretation and implementation of this tracer methodology, coping with a large team of researchers, obtaining ethics approval and establishing the governance structure, deployment of the team to the site, ensuring consistency in the data collection, managing data sets, and the reliability of the coding. The methodology adopted is time-consuming, and the very large data sets that are generated need to be managed. Complex research project management structures and processes are required, which would not be needed for traditional cross-sectional studies. Sufficient time needs to be allowed at the start of such research projects, in order to put the necessary systems in place. The paper will be of interest to OSH researchers and those contemplating longitudinal studies, particularly those employing a tracer approach.