Description
Traces incidents through layered defences to show how latent conditions and active failures align, so design fixes target system weaknesses rather than blaming users.
Reason's model represents a system's safeguards as slices of Swiss cheese, each with holes caused by active failures (unsafe acts by people at the sharp end) and latent conditions (design decisions, staffing, procedures). Harm occurs when holes in successive layers align. As a research method it structures incident analysis: for each incident or near miss, analysts reconstruct the sequence, identify every defence that should have stopped it, and classify why each failed. The resulting picture usually shows that interface design, alerts and confirmations are some of the layers, and that user error is an outcome rather than a root cause. Rigour comes from systematic reconstruction with evidence, analysing near misses as well as harms, and aggregating across incidents to find recurring holes. Frameworks such as HFACS formalise the classification.
Attributes
- Discipline
- Human Factors & Error
- Loss type
- Siloed
- Stage
- Post-launch
- Data
- Qualitative
- Effort
- Weeks
Steps
- Collect incidents and near misses with supporting evidence such as logs, screenshots and interviews.
- Reconstruct each event timeline from initiating action to outcome.
- List every defence layer that could have intercepted it: interface constraints, confirmations, alerts, procedures, reviews.
- Classify each layer's failure as an active failure or latent condition, using a scheme such as HFACS.
- Aggregate across incidents to find layers that fail repeatedly.
- Design fixes that strengthen recurrent weak layers or add independent ones, and monitor near-miss rates.
Sources
- Handbook of Human Factors and Ergonomics
- Human Error (James Reason)
- Managing the Risks of Organizational Accidents (James Reason)
Also known as
- Reason's Swiss Cheese Model
- Organisational accident model
- Defences-in-depth analysis
Categories
- Accessibility
- Human Factors & Error