CAA NZ AME Human Factors cheat sheet

Human factors is the study of how people interact with their work environment; in maintenance its purpose is to reduce error and protect airworthiness, not to assign blame.

The 'dirty dozen' error precursors are: fatigue, complacency, distraction, lack of communication, lack of teamwork, lack of resources, pressure, lack of assertiveness, stress, lack of awareness, norms, and lack of knowledge.

The SHEL model frames mismatch at interfaces: Software (procedures, manuals), Hardware (tools, aircraft), Environment (light, noise, temperature), and Liveware (people, including self and others).

Reason's Swiss-cheese model: latent conditions (organisational, design, regulatory) create holes in defences; active failures by front-line staff pass through when holes align.

Error types: slips (correct intention, wrong execution), lapses (memory failure), mistakes (wrong intention: rule-based or knowledge-based), and violations (deliberate deviation from procedure).

Maintenance error decision aid (MEDA) logic: identify the error, then the contributing factors (individual, supervision, organisation), then corrective actions; it targets system causes rather than individuals.

The 'dirty dozen' are symptoms of system conditions, so the corrective action should change the system (e.g. task lighting, dual inspection, clearer documentation) rather than retrain the individual alone.

Error rates rise with task complexity and interruption; a second set of eyes (independent inspection) is a defence against slips and lapses, not against violations.

Fatigue degrades vigilance, reaction time, and decision quality; it is a normal physiological state, not a character flaw, and is managed by rest, rostering and workload control.

Circadian low points typically occur in the early hours of the morning (roughly 02:00-06:00) and mid-afternoon; error risk is highest then, especially on night shifts.

Acute stress narrows attention and can speed up work unsafely; chronic stress impairs memory and judgement and is linked to absenteeism and error.

The Yerkes-Dodson curve: performance improves with arousal up to an optimum, then declines; complex tasks have a lower optimum arousal than simple tasks.

Alcohol, some medications, illness and sleep debt all increase fatigue; the AME should declare fitness-to-work concerns rather than self-medicate.

Fatigue risk management uses limits on duty hours, minimum rest, controlled napping, and fatigue reporting; it is a shared responsibility between the individual and the organisation.

Stressors in maintenance include noise, heat, vibration, time pressure, shift work, and interpersonal conflict; recognising the stressor is the first step to mitigating it.

Effective communication requires a sender, a message, a medium, and a receiver; feedback confirms the message was received as intended, and without feedback communication is unverified.

Shift handover should be structured (e.g. written log plus verbal briefing plus confirmation), covering open tasks, deferred defects, and outstanding inspections.

Assertiveness is the ability to state a safety concern clearly and respectfully; it is a defence against the 'authority gradient' where a junior person defers to a senior.

A steep authority gradient suppresses challenge; a flat one can cause confusion; the safest teams have clear roles but encourage challenge.

Teamwork requires shared mental models, clear roles, and mutual monitoring; a team that does not cross-check each other is more error-prone.

Barriers to communication include noise, jargon, language differences, assumptions, and 'normalisation of deviance' where unsafe practice becomes accepted.

Written communication (work cards, tech logs) must be legible, unambiguous, and complete; ambiguous sign-offs are a recurring maintenance error source.

Situational awareness has three levels: perception of elements, comprehension of their meaning, and projection of future status; loss of SA is a precursor to most errors.

Loss of SA causes include task saturation, fatigue, distraction, expectation bias ('I expected it to be that way'), and fixation on one task.

Decision-making models (e.g. DECIDE, FOR-DEC) structure a choice: define the problem, gather information, consider options, decide, act, and evaluate.

Confirmation bias is the tendency to seek evidence that supports an existing belief; it is countered by actively seeking disconfirming evidence and independent checks.

Complacency is a state of reduced vigilance after repeated success; it is countered by deliberate checks, checklists, and rotating attention.

A 'just culture' distinguishes human error (console), at-risk behaviour (coach), and reckless behaviour (sanction); this encourages reporting and learning.

Human error reporting systems (e.g. MOR, internal occurrence reports) feed safety management; reporting is protected only when the reporter is not reckless.

Procedures and work cards are a primary defence against error; they must be current, clear, and matched to the task, and deviations must be authorised and recorded.

The technical log records defects, deferred items, and certifications; entries must be accurate, signed, and traceable to the person and licence.

Safety management systems (SMS) have four components: safety policy, safety risk management, safety assurance, and safety promotion.

Risk assessment uses likelihood x severity matrices; risk is reduced by eliminating, substituting, engineering, administrative controls, or PPE, in that order of preference.

A safety culture has reporting, just, flexible, and learning dimensions; a blame culture suppresses reporting and hides risk.

Human factors training is a regulatory requirement in maintenance organisations; it should be recurrent and integrated with occurrence data.

Audits and occurrence investigations look for systemic causes; the '5 whys' and fishbone diagrams help trace an error back to its root cause.

Environmental stressors in maintenance include noise (impairs communication and concentration), heat and cold (impair dexterity and vigilance), vibration, poor lighting, and confined spaces.

Illumination must suit the task; inadequate or glare-prone lighting increases inspection errors and is a common contributor to missed defects.

Workload has an inverted-U relationship with performance; both underload (boredom, complacency) and overload (task saturation) raise error rates.

Task saturation occurs when demand exceeds capacity; it is managed by prioritising, delegating, and stopping to reassess rather than rushing.

Physical factors: vibration and repetitive tasks cause fatigue and injury; ergonomics of tool and workspace design affects error and health.

Shift work and night duty disrupt circadian rhythm; scheduling of complex tasks should avoid circadian low points where practicable.

Noise and communication: high noise forces shouting and mishearing; hearing protection and clear verbal protocols reduce error.

Inspection is a vigilance task: performance declines over time (vigilance decrement), so inspections should be time-limited with breaks and rotated staff.

Expectation and set bias cause inspectors to see what they expect; independent inspection and varied inspection routes counter this.

Nondestructive testing (NDT) reliability depends on technique, training, and human factors; a single inspector's judgement is not infallible.

The 'look-alike' problem: similar components or connectors can be fitted incorrectly; colour coding, keying, and part marking are defences.

Post-maintenance verification (functional check, independent inspection) catches slips and lapses before release to service.

The 'dirty dozen' apply directly to inspection: fatigue, complacency, distraction, and lack of awareness are the most common causes of missed defects.

Recording findings accurately and reporting near-misses builds the data needed to improve inspection reliability.

Human factors is embedded in maintenance organisation approvals; the organisation must have procedures for error reporting, investigation, and corrective action.

The regulator (CAA NZ) requires occurrence reporting; the purpose is safety learning, not punishment, and reports feed the safety management system.

Just culture: honest error should be consoled and learned from; reckless or deliberate violation may be sanctioned; the boundary is the key judgement.

Human factors training records must be maintained and available for audit; recurrent training is typically required.

Maintenance error investigation should identify the error type (slip, lapse, mistake, violation) and the contributing factors before selecting corrective action.

Safety data is protected when reported through the correct channel; the reporter should not alter or conceal evidence.

Regulatory oversight uses audits, inspections, and occurrence data; a positive safety culture is assessed through reporting rates and follow-up actions.

SHEL: Software, Hardware, Environment, Liveware; mismatches at the Liveware-Liveware interface are often the most safety-critical.

Dirty dozen: fatigue, complacency, distraction, lack of communication, lack of teamwork, lack of resources, pressure, lack of assertiveness, stress, lack of awareness, norms, lack of knowledge.

Swiss cheese: latent conditions plus active failures align to breach defences; defences include procedures, training, inspection, and design.

Reason's error types: slips and lapses (execution), mistakes (planning), violations (intentional deviation).

DECIDE / FOR-DEC: structured decision-making models for time-pressured choices.

Yerkes-Dodson: arousal-performance inverted U; complex tasks need lower arousal.

Just culture: console human error, coach at-risk behaviour, sanction recklessness.

If a task is interrupted, return to the start of the step and verify before continuing; do not rely on memory of where you were.

If you are fatigued, do not perform safety-critical tasks; report fatigue and seek relief rather than pushing through.

If you are unsure, stop and ask; a question costs minutes, an error can cost lives.

Sign only for work you personally performed or directly supervised and verified; never sign for another person's work.

Use the work card as the authority; if the card is wrong, raise a discrepancy and get it corrected rather than improvising.

Independent inspection must be genuinely independent: a different person, a different route, and no prompting of the expected answer.

Report near-misses and errors promptly; the report is the raw material for system improvement.

Use closed-loop communication: sender states message, receiver repeats back, sender confirms; this catches misheard instructions.

Standard phraseology and plain language reduce ambiguity; avoid slang and assumptions about shared knowledge.

Handover must include: task status, open defects, parts removed, tools used, and any deferred work; confirm understanding with the incoming person.

Written records must be contemporaneous; memory fades and retrospective entries are unreliable and non-compliant.

Escalate safety concerns through the chain of command; if the immediate supervisor does not act, go higher or use the reporting system.

Brief the team before a complex task: roles, sequence, hazards, and stop criteria; this builds a shared mental model.

Start with what happened, not who to blame; establish the sequence of events and the task conditions.

Classify the error: slip, lapse, mistake (rule- or knowledge-based), or violation; the classification guides the fix.

Identify contributing factors across the SHEL interfaces and the dirty dozen; look for system conditions, not just individual acts.

Select corrective actions that change the system: redesign tasks, improve documentation, add independent checks, adjust rostering, improve lighting.

Verify corrective actions are implemented and effective; track recurrence through occurrence data.

Share lessons learned across the organisation; a single investigation should improve practice beyond the individual concerned.

Fitness to work means being physically and mentally able to perform the task safely; illness, fatigue, alcohol, drugs, and stress can all impair fitness.

Alcohol: even small amounts impair judgement and reaction time; the safe rule is not to work under the influence, and to allow time for clearance.

Medication: some over-the-counter and prescription drugs cause drowsiness; check with an AME or medical adviser before safety-critical work.

Stress and personal problems can distract; use support systems and report if you are not fit.

Sleep: aim for adequate sleep before duty; fatigue is cumulative and cannot be fully reversed by one short break.

Self-declaration of unfitness is a professional duty; hiding unfitness is a violation and a safety risk.

Safety culture dimensions: reporting, just, flexible, and learning; each supports the others.

A reporting culture depends on trust; punishment for honest error destroys reporting and hides risk.

A just culture distinguishes error, at-risk behaviour, and recklessness; the response should match the behaviour.

A flexible culture adapts to change and unexpected events; rigid adherence to routine can fail in novel situations.

A learning culture uses occurrence data, audits, and training to improve; lessons must be fed back to the workforce.

Leadership sets the tone: supervisors who prioritise schedule over safety create pressure that leads to violations.

Norms (the 'dirty dozen') are unwritten rules; they can normalise unsafe practice and must be challenged by leadership and peers.

Time pressure is a dirty-dozen item; it narrows attention and encourages shortcuts and violations.

Under-resourcing (staff, parts, tools, time) is a system condition that raises error; it must be reported and managed, not absorbed silently.

Task prioritisation should be based on safety criticality, not convenience; defer non-critical work rather than rush critical work.

Workload monitoring: if demand exceeds capacity, stop and reassess; task saturation is a known precursor to error.

Pressure can be commercial, supervisory, or self-imposed; recognising the source helps in resisting it.

Adequate resources include correct tooling, current manuals, and sufficient trained staff; lack of any is a reportable condition.

Distraction is a leading cause of maintenance error; interruptions during a task should be minimised and the task re-verified on return.

Complacency follows repeated success; it reduces vigilance and is countered by deliberate checks and rotating attention.

Lack of awareness means not noticing a change or hazard; it is countered by briefings, checklists, and cross-checks.

Norms are unwritten rules that can normalise unsafe practice; challenge them when they conflict with procedures.

Lack of knowledge is a dirty-dozen item; it is addressed by training, mentoring, and using current documentation.

Lack of assertiveness means not speaking up; it is countered by a culture that welcomes challenge and by practising clear statements.

Lack of teamwork means poor coordination; it is countered by briefings, shared mental models, and mutual monitoring.

Attention is limited; multitasking degrades performance, so tasks should be sequenced and interruptions minimised.

Memory is fallible; reliance on memory for procedures is unsafe, so work cards and checklists are used.

Perception is subject to illusion and expectation; independent verification counters perceptual error.

Decision-making is affected by bias (confirmation, anchoring, availability); structured models and challenge reduce bias.

Physical limits: strength, reach, and dexterity; ergonomics and tooling design must match the person.

Stress and fatigue degrade all of the above; managing them is a core human factors skill.

Report the occurrence promptly through the organisation's system; include what, when, where, and the conditions.

Preserve evidence (parts, photos, records) before it is disturbed; investigation depends on it.

Investigation should be independent of the people involved where practicable; this protects objectivity.

Use structured methods (5 whys, fishbone, MEDA) to trace from the error to contributing factors.

Corrective actions should be specific, assigned, and time-bound; verify effectiveness.

Feed lessons back to the workforce and into training; this closes the learning loop.

Protect confidentiality where the system allows; trust is essential for future reporting.

A slip is a correct intention executed incorrectly (e.g. wrong switch); a lapse is a memory failure (e.g. missed step).

A mistake is a correct execution of a wrong plan; it can be rule-based (misapplied rule) or knowledge-based (lack of knowledge).

A violation is a deliberate deviation from a procedure; it may be routine (accepted practice) or exceptional (situation-driven).

Violations are often caused by procedures that are impractical or by pressure; the fix may be to change the procedure, not just punish.

Error management focuses on containing the consequences of error, not only on preventing it.

Defences against error include design (error-proofing), procedures, training, and independent checks.

Just culture response: console error, coach at-risk behaviour, sanction recklessness.

Briefings before complex tasks align the team on roles, sequence, hazards, and stop criteria.

Closed-loop communication (repeat-back) confirms understanding and catches misheard instructions.

Written records must be legible, complete, and contemporaneous; ambiguous entries are a safety risk.

Handover must cover open tasks, deferred defects, and outstanding inspections, with confirmation.

Escalate concerns through the chain of command; if not acted on, use the reporting system.

Language and jargon barriers should be managed with plain language and confirmation of understanding.

Assertiveness: state the concern, the reason, and the requested action; do not soften a safety concern into a suggestion.

SMS components: safety policy, safety risk management, safety assurance, safety promotion.

Risk = likelihood x severity; assess before and after controls to confirm residual risk is acceptable.

Control hierarchy: eliminate, substitute, engineering controls, administrative controls, PPE.

Safety assurance uses audits, occurrence data, and performance indicators to confirm controls work.

Safety promotion includes training, communication, and a positive culture.

Human factors training is part of safety promotion and should be recurrent.

Occurrence data feeds risk management; reporting is the source of that data.

If interrupted, restart the step and verify; do not rely on memory.

If fatigued, do not perform safety-critical tasks; report and seek relief.

If unsure, stop and ask; a question is cheaper than an error.

Sign only for work you performed or directly supervised and verified.

Use the work card as authority; raise a discrepancy if it is wrong rather than improvising.

Independent inspection must be genuinely independent: different person, different route, no prompting.

Report near-misses and errors promptly; they are the raw material for improvement.

SHEL model: Software, Hardware, Environment, Liveware; mismatch at interfaces causes error.

Swiss cheese: latent conditions and active failures align to breach defences.

Dirty dozen: fatigue, complacency, distraction, lack of communication, lack of teamwork, lack of resources, pressure, lack of assertiveness, stress, lack of awareness, norms, lack of knowledge.

Error types: slip, lapse, mistake (rule/knowledge), violation.

Yerkes-Dodson: arousal-performance inverted U; complex tasks need lower arousal.

DECIDE/FOR-DEC: structured decision models.

Just culture: console error, coach at-risk behaviour, sanction recklessness.

Fatigue degrades vigilance, reaction time, and decision quality; it is a physiological state managed by rest and rostering.

Circadian low points: early morning and mid-afternoon; error risk is highest then.

Acute stress narrows attention; chronic stress impairs memory and judgement.

Yerkes-Dodson: performance peaks at optimum arousal; complex tasks have a lower optimum.

Alcohol, medication, illness, and sleep debt increase fatigue; declare unfitness rather than self-medicate.

Fatigue risk management: duty limits, minimum rest, controlled napping, reporting.

Stressors: noise, heat, vibration, time pressure, shift work, conflict; identify and mitigate.

Effective communication needs sender, message, medium, receiver, and feedback.

Structured handover: written log plus verbal briefing plus confirmation.

Assertiveness states a safety concern clearly and respectfully; it counters the authority gradient.

Authority gradient: steep suppresses challenge; flat can confuse; clear roles plus encouraged challenge is safest.

Teamwork needs shared mental models, clear roles, and mutual monitoring.

Barriers: noise, jargon, language, assumptions, normalisation of deviance.

Written communication must be legible, unambiguous, and complete.

SA levels: perception, comprehension, projection; loss of SA precedes most errors.

Causes of SA loss: task saturation, fatigue, distraction, expectation bias, fixation.

Decision models: DECIDE, FOR-DEC; define, gather, consider, decide, act, evaluate.

Confirmation bias: seek disconfirming evidence and independent checks.

Complacency: reduced vigilance after repeated success; counter with deliberate checks.

Just culture: console error, coach at-risk behaviour, sanction recklessness.

Occurrence reporting feeds safety management and learning.

Procedures and work cards are a primary defence; they must be current, clear, and matched to the task.

Deviations must be authorised and recorded; improvisation is a violation risk.

Technical log entries must be accurate, signed, and traceable.

SMS: policy, risk management, assurance, promotion.

Risk control hierarchy: eliminate, substitute, engineering, administrative, PPE.

Safety culture: reporting, just, flexible, learning.

Investigation uses 5 whys and fishbone to find root causes.

Environmental stressors: noise, heat, cold, vibration, poor lighting, confined spaces.

Illumination must suit the task; glare and poor light increase inspection errors.

Workload-performance is an inverted U; underload and overload both raise error.

Task saturation: demand exceeds capacity; prioritise, delegate, stop and reassess.

Vibration and repetition cause fatigue and injury; ergonomics matters.

Shift work disrupts circadian rhythm; schedule complex tasks away from low points.

Noise impairs communication; hearing protection and clear protocols reduce error.

Inspection is a vigilance task; performance declines over time (vigilance decrement).

Time-limit inspections with breaks and rotate staff.

Expectation and set bias cause inspectors to see what they expect; independent inspection counters this.

NDT reliability depends on technique, training, and human factors.

Look-alike components can be fitted incorrectly; colour coding, keying, and marking are defences.

Post-maintenance verification catches slips and lapses before release.

Dirty dozen apply to inspection: fatigue, complacency, distraction, lack of awareness.

Human factors is embedded in maintenance organisation approvals; procedures for error reporting and corrective action are required.

CAA NZ requires occurrence reporting; purpose is safety learning, not punishment.

Just culture: honest error consoled; reckless violation may be sanctioned.

Human factors training records must be maintained and available for audit.

Error investigation identifies error type and contributing factors before corrective action.

Safety data is protected when reported through the correct channel.

Oversight uses audits, inspections, and occurrence data.

SHEL: Software, Hardware, Environment, Liveware.

Dirty dozen: fatigue, complacency, distraction, lack of communication, lack of teamwork, lack of resources, pressure, lack of assertiveness, stress, lack of awareness, norms, lack of knowledge.

Swiss cheese: latent conditions plus active failures breach defences.

Reason's error types: slip, lapse, mistake, violation.

DECIDE / FOR-DEC: structured decisions.

Yerkes-Dodson: inverted U arousal-performance.

Just culture: console, coach, sanction.

If interrupted, restart the step and verify.

If fatigued, do not perform safety-critical tasks; report and seek relief.

If unsure, stop and ask.

Sign only for work you performed or directly supervised and verified.

Use the work card as authority; raise a discrepancy if wrong.

Independent inspection must be genuinely independent.

Report near-misses and errors promptly.

Closed-loop communication: sender states, receiver repeats, sender confirms.

Standard phraseology and plain language reduce ambiguity.

Handover includes task status, open defects, parts removed, tools used, deferred work.

Written records must be contemporaneous.

Escalate safety concerns through the chain of command.

Brief the team before complex tasks: roles, sequence, hazards, stop criteria.

Start with what happened, not who to blame.

Classify the error: slip, lapse, mistake, violation.

Identify contributing factors across SHEL and the dirty dozen.

Select system-level corrective actions.

Verify actions are implemented and effective.

Share lessons learned across the organisation.

Fitness to work: physically and mentally able to perform safely.

Alcohol impairs judgement; do not work under the influence.

Medication can cause drowsiness; check before safety-critical work.

Stress and personal problems distract; use support systems.

Sleep debt is cumulative; one short break does not reverse it.

Self-declaration of unfitness is a professional duty.

Safety culture dimensions: reporting, just, flexible, learning.

Reporting depends on trust; punishment for honest error destroys reporting.

Just culture distinguishes error, at-risk behaviour, recklessness.

Flexible culture adapts to change and novel situations.

Learning culture uses data, audits, training, and feedback.

Leadership sets the tone; schedule pressure can create violations.

Norms are unwritten rules; challenge unsafe ones.

Time pressure narrows attention and encourages shortcuts.

Under-resourcing raises error; report and manage it.

Prioritise by safety criticality, not convenience.

If demand exceeds capacity, stop and reassess.

Pressure can be commercial, supervisory, or self-imposed.

Adequate resources include tooling, manuals, and trained staff.

Distraction is a leading cause of maintenance error; minimise interruptions.

Complacency follows repeated success; counter with deliberate checks.

Lack of awareness: not noticing a change or hazard; counter with briefings and cross-checks.

Norms can normalise unsafe practice; challenge them.

Lack of knowledge: address with training and current documentation.

Lack of assertiveness: practise clear statements and support challenge.

Lack of teamwork: briefings, shared mental models, mutual monitoring.

Attention is limited; multitasking degrades performance.

Memory is fallible; use work cards and checklists.

Perception is subject to illusion and expectation; verify independently.

Decision-making is affected by bias; use structured models and challenge.

Physical limits: strength, reach, dexterity; ergonomics matters.

Stress and fatigue degrade all performance; manage them.

Report promptly through the organisation's system.

Preserve evidence before it is disturbed.

Investigation should be independent where practicable.

Use structured methods: 5 whys, fishbone, MEDA.

Corrective actions must be specific, assigned, and time-bound.

Feed lessons back to the workforce and training.

Protect confidentiality where allowed.

Slip: correct intention, wrong execution.

Lapse: memory failure.

Mistake: correct execution of wrong plan; rule- or knowledge-based.

Violation: deliberate deviation; routine or exceptional.

Violations often stem from impractical procedures or pressure.

Error management contains consequences as well as preventing error.

Defences: design, procedures, training, independent checks.

Briefings align the team on roles, sequence, hazards, stop criteria.

Closed-loop communication confirms understanding.

Written records must be legible, complete, contemporaneous.

Handover covers open tasks, deferred defects, outstanding inspections.

Escalate concerns through the chain of command.

Manage language and jargon barriers with plain language.

Assertiveness: state concern, reason, requested action.

SMS: policy, risk management, assurance, promotion.

Risk = likelihood x severity; assess before and after controls.

Control hierarchy: eliminate, substitute, engineering, administrative, PPE.

Safety assurance uses audits, occurrence data, indicators.

Safety promotion includes training, communication, culture.

Human factors training is part of safety promotion.

Occurrence data feeds risk management.

If interrupted, restart the step and verify.

If fatigued, do not perform safety-critical tasks.

If unsure, stop and ask.

Sign only for work you performed or supervised and verified.

Use the work card as authority; raise a discrepancy if wrong.

Independent inspection must be genuinely independent.

Report near-misses and errors promptly.