Human factors & safety culture

040 Human PerformanceChapter 1

Aviation is extremely safe, but human factors remain the dominant cause when things go wrong. Understanding error models helps design safer operations.

Safety statistics

Key fact

Aviation averages about 1 accident per million movements — the safest mass transport mode. GPWS has been the largest recent contributor to safety improvement.

Key fact

Approximately 70–80% of accidents involve human factors — often poor judgement, not lack of technical skill.

Key fact

Realistic human error rates: ~1 in 1000 for skilled tasks; simple repetitive tasks fail about 1 in 100.

Error types

Type	Definition	Example
Fault	Wrong intention, action seems correct	Shutting down the wrong engine
Slip	Correct intention, wrong action	Selecting flaps instead of gear
Omission	Forgetting to act	Missing checklist item
Commission	Doing what should not be done	Takeoff without clearance
Substitution	Wrong action substituted	Similar to slip

Definition — Active error

Has immediate consequence — visible at the sharp end (pilot action).

Definition — Latent error

Built into the system — poor cockpit design, bad procedures, organisational pressure. Example: cockpit design error as latent threat.

Exam trap

Fault = wrong intention. Slip = right intention, wrong execution. Exam questions often swap these definitions.

Error tolerance & culture

Definition — Error tolerance

System design where no single error can cause failure — defences absorb mistakes.

Key fact

Safety culture is a subset of national culture. An open culture encourages reporting; a closed culture suppresses it due to fear or lack of incentive.

Key fact

James Reason’s informed culture: gather data from accidents and investigations, assess reports, learn, and implement flexible solutions. Just culture — genuine mistakes are not punished.

Swiss cheese model

Schematic

Multiple defence layers — accident occurs when active and latent failures align through all holes.

Key fact

Accidents occur when active and latent failures align through successive layers of defences (like holes in cheese slices lining up).

Key fact

Typically 4 to 7 links in an error chain lead to an accident.

SHEL model

Key fact

SHEL(L) — mismatch between system components causes errors:

• Software — procedures, checklists, manuals
• Hardware — aircraft, controls, displays
• Environment — weather, noise, temperature
• Liveware — people (pilot × pilot, pilot × other liveware)

Key fact

Liveware–software mismatch is a common conceptual systems mismatch (e.g. procedure not suited to human capability).

Threat & Error Management (TEM)

Schematic

Threats (external) lead to errors (crew) and potentially an undesired aircraft state — countered by hard and soft defences.

Definition — Threat

External factor beyond crew control — weather, ATC, organisational pressure.

Definition — Error

Crew action or inaction — procedural, communication, aircraft handling.

Definition — Undesired Aircraft State (UAS)

Unintended situation reducing safety margin — unstable approach, wrong config.

Key fact

Hard countermeasures — built into system before flight (TCAS, GPWS, checklists). Soft countermeasures — crew KSA during flight (CRM, cross-check).

Operation modes (Rasmussen / Reason)

Mode	Description	Error type
Skill-based	Automatic, routine	Slips
Rule-based	If–then procedures	Rule errors
Knowledge-based	Novel problem solving	Knowledge errors

Key fact

A competent pilot has correct Knowledge, Skills and Attitudes (KSA). Everyone is responsible for safety — not only the captain.