UCL School of Management is delighted to welcome John Caroll, MIT, to host a research seminar discussing ‘Adventures in Safety Culture: How Mental Models Affect Safety in High-Hazard Industries’​

Abstract

“Safety culture” was invented about 30 years ago as an explanation for the Chernobyl accident. Something more than defective equipment and human error seemed to underlie the surprising events that led to the worst accident in nuclear power history (until Fukushima, perhaps). Despite considerable efforts in research, policy, consulting, and training, the challenges of managing culture have not dissipated over time. Increasingly, diverse industries and multiple countries are mandating safety culture (or climate) measurement and improvement as part of the regulatory process, but aside from institutional legitimation and impression management, the tangible benefits of safety culture management (i.e. improving safety) are contested. I argue in this presentation that mental models about safety are themselves diverse and these frequently unrecognized and undiscussed differences are an impediment to effective action to improve safety.  By “mental models”, I include sensemaking and causal beliefs on the part of managers, workers, safety professionals, regulators, researchers, and others about the causes of accidents and the mechanisms of effective safety management, including safety culture as a cause of good or bad safety performance.  Mental models underlie expectations about safety culture, its meaning, measurement, and management. Mental models underlie our understanding of accidents and errors (individual and collective) and the processes through which accidents are investigated and changes imagined and implemented.  Mental models are themselves a component of culture, as an aspect of shared sensemaking and a wellspring of action planning although, like culture, they may be widely shared or locally (even individually) held. For example, studies of “safety” rarely distinguish personal or industrial safety (exemplified by slips, trips, and falls and measured by days away from work and similar metrics) from process or system safety (that underlies rare but severe accidents with a confluence of causes, conditions, and triggers); as a result, research unintentionally subsumes system safety under personal safety because personal safety events are much more common and visible. I draw on theories from cognitive social psychology and organizational behavior along with qualitative and quantitative data from several high hazard industries to offer an agenda for research connecting mental models, culture management and safety improvement.