Can we examine safety culture in accident investigations, or should we?

[Note: this is an update and repost of a really old summary which had a different format back then. I skipped a huge amount of the author's arguments in this paper, so you'll need to read the full paper to get most of the background info and context.]

Abstract

Considerable attention has been paid to safety culture since the 1986 Chernobyl nuclear power station accident. Researchers have studied it and companies and regulators have applied it to enhance safety. However, few research studies have been conducted that establish a link between safety culture and operational or process safety and methods used to assess safety culture, primarily questionnaires, have been criticized on methodological grounds. One way to enhance system safety is through applying the lessons of investigations of accidents of process safety to remediate organizational shortcomings identified in the investigation. Rather than attempting to assess safety culture directly, examining company actions and decisions directly after an accident can allow investigators to make inferences about safety culture at the time of the accident. This study suggests a method to directly examine the role of organizations in accidents by identifying the nature of organizational errors and describing the logic that can link these errors to accident causation. The application of this method in several accident investigations is described.

****

From the full-text article:

·        In answering whether we can examine safety culture in accident investigations or whether we should, the author argues that “Directly assessing a construct such as safety culture after the fact of an accident, in an effort to gauge its state before an accident, is not supported by the findings of research and of accident investigations” (pg. 109);

·        “the link between safety culture and process safety is only suggestive as most of the research into such a relationship has been largely based on measures of occupational safety and safety climate, not process safety and safety culture. As the Chemical Safety and Hazard Investigation Board observed, focusing on occupational safety does not necessarily relate to enhanced operational or process safety” (pg. 109);

·        “limitations in our understanding and measurement of safety culture preclude direct assessment of safety culture in accident investigations.” (pg. 109), and “both company managers and operators may well believe that their processes are safe, beliefs that would be expected to affect the results of a direct assessment of safety culture, only to discover that the contrary is true from the results of an investigation” (pg. 109);

·        The author suggests that company practices as identified during an investigation, such as indicating how work is actually done, perceptions of hazards, use and value of procedures etc, can indicate more about its safety practices about safety culture than can perhaps safety climate/culture data (with the exception of extensive ethnographic study); and

·        The author also suggests that further research is needed examining the difference between perceived safety and actual safety, posing the question of how managers can recognise whether their beliefs in operational safety of their systems are accurate and how they can identify flaws in operational safety.

He further argues that even if companies undertake a range of initiatives to build, enhance or change “cultural elements”, based on his comparison to the two accidents in the paper these elements may have little direct connection to operational safety.

Paraphrasing the author, he says that focusing on some cultural elements may give managers a false sense of safety which could be disconnected from operational safety. However, he further says that this is not to say that we shouldn’t try to understand culture nor denigrate its importance.

Rather “directly assessing safety culture in an accident investigation can be ineffective or worse, misleading” and that the information an organisation normally collects in routine operations, and learning activities and investigations can “better describe a company’s actual practices in system operations than could be obtained from most direct assessments of safety culture”.

 Link in comments.

Strauch, B. (2015). Can we examine safety culture in accident investigations, or should we? Safety Science, 77, 102-111.