In workplaces where activity continues around the clock, human error has been
observed to exhibit a circadian rhythm, with a characteristic peak in the early hours
of the morning. Errors are commonly distinguished by the nature of the underlying
cognitive failure, particularly the level of intentionality involved in the erroneous
action. The Skill-Rule-Knowledge (SRK) framework of Rasmussen is used widely in
the study of industrial errors and accidents. The SRK framework describes three fundamental
types of error, according to whether behavior is under the control of practiced
sensori-motor skill routines with minimal conscious awareness; is guided by
implicit or explicit rules or expertise; or where the planning of actions requires the
conscious application of domain knowledge. Up to now, examinations of circadian
patterns of industrial errors have not distinguished between different types of error.
Consequently, it is not clear whether all types of error exhibit the same circadian
rhythm. A survey was distributed to aircraft maintenance personnel in Australia.
Personnel were invited to anonymously report a safety incident and were prompted
to describe, in detail, the human involvement (if any) that contributed to it. A total of
402 airline maintenance personnel reported an incident, providing 369 descriptions
of human error in which the time of the incident was reported and sufficient detail
was available to analyze the error. Errors were categorized using a modified version
of the SRK framework, in which errors are categorized as skill-based, rule-based, or
knowledge-based, or as procedure violations. An independent check confirmed that
the SRK framework had been applied with sufficient consistency and reliability. Skillbased
errors were the most common form of error, followed by procedure violations,
rule-based errors, and knowledge-based errors. The frequency of errors was adjusted
for the estimated proportion of workers present at work/each hour of the day, and
the 24 h pattern of each error type was examined. Skill-based errors exhibited a significant
circadian rhythm, being most prevalent in the early hours of the morning.
Variation in the frequency of rule-based errors, knowledge-based errors, and procedure
violations over the 24 h did not reach statistical significance. The results
suggest that during the early hours of the morning, maintenance technicians are at