Human Factors and Decision Making

Why Human Factors and Decision Making Matter

Most serious incidents are not caused by a single broken part or a careless worker. Investigations across aviation, oil and gas, healthcare, and manufacturing consistently show that 70 to 90 percent of incidents involve a human performance component: a misread gauge, a skipped step under time pressure, a decision made with incomplete information, or a design that invited the wrong action. The Associate Safety Professional (ASP) exam places Human Factors within the Training, Education, and Communication domain because a safety professional cannot simply tell people to be more careful. You must understand why capable, well-intentioned people make errors and poor decisions, and then design systems, training, and communication that make the right action the easy action. On the exam, this topic tests whether you can move beyond blaming the individual to identifying the latent conditions, cognitive limits, and organizational pressures that shape behavior.

What Human Factors Is

Human factors (also called ergonomics in its broad sense) is the scientific discipline concerned with understanding interactions among humans and other elements of a system, and applying that knowledge to optimize human well-being and overall system performance. For the ASP exam, think of it in three layers:

1. Physical human factors: anthropometry, biomechanics, reach, strength, fatigue, lighting, noise, and workstation design.
2. Cognitive human factors: perception, attention, memory, workload, situational awareness, and decision making.
3. Organizational human factors: culture, communication, staffing, shift schedules, production pressure, and management systems.

The exam heavily emphasizes the cognitive and organizational layers when it uses the phrase human factors and decision making.

Key Models and Frameworks You Must Know

1. The SHELL Model
Developed in aviation, SHELL describes the human (Liveware, at the center) interacting with Software (procedures, checklists, symbology), Hardware (equipment, controls, displays), Environment (lighting, temperature, noise, time pressure), and other Liveware (coworkers, supervisors, communication). Mismatches at any interface produce error. Exam cue: if a question describes a control that looks identical to another control and gets confused, that is a Liveware-Hardware mismatch.

2. James Reason's Swiss Cheese Model
Defenses are layered like slices of cheese, each with holes. Active failures are unsafe acts by front-line workers with immediate consequences. Latent conditions are decisions made upstream (design, budget, scheduling, training gaps) that lie dormant until they align with an active failure. The exam loves to ask you to distinguish active failures from latent conditions. Rule of thumb: if it was decided in a conference room months ago, it is latent; if it happened at the point of work today, it is active.

3. Reason's Classification of Unsafe Acts
Memorize these four categories and their definitions:
- Slips: attention failures during a routine, skill-based task. The intention was correct but the execution was wrong (pressing the wrong button while distracted).
- Lapses: memory failures, usually omissions (forgetting a step in a familiar procedure).
- Mistakes: planning failures. The action went as intended, but the plan was wrong. These divide into rule-based mistakes (applying the wrong rule or misapplying a good rule) and knowledge-based mistakes (reasoning incorrectly in a novel situation).
- Violations: deliberate deviations from rules or procedures. These may be routine (habitual corner-cutting), situational (the rule cannot be followed under current conditions), optimizing (for thrill or personal gain), or exceptional (one-off response to an unusual situation).
Critical distinction: slips and lapses are errors of execution; mistakes are errors of intention; violations are intentional departures. The correct corrective action differs for each, and the exam will test that.

4. Rasmussen's Skill-Rule-Knowledge (SRK) Framework
- Skill-based behavior: highly practiced, automatic actions requiring little conscious thought (an experienced driver shifting gears). Errors here are slips and lapses. Error rate is low (roughly 1 in 1,000) but the person is not actively thinking about the task.
- Rule-based behavior: recognizing a situation and applying a stored if-then rule or procedure. Errors are misapplied rules. Error rate is moderate (roughly 1 in 100).
- Knowledge-based behavior: novel situations requiring conscious problem solving from first principles. Error rate is high (roughly 1 in 2 to 1 in 10) because working memory is limited and the person is improvising.
Exam application: training should aim to move critical tasks from knowledge-based toward rule-based (with procedures and job aids) and ultimately toward skill-based performance through practice and drills, while adding verification steps to catch skill-based slips.

5. Situational Awareness (Endsley's Three Levels)
- Level 1: Perception of elements in the environment.
- Level 2: Comprehension of what those elements mean.
- Level 3: Projection of their future status.
Most situational awareness failures occur at Level 1 (the information was available but not noticed). Loss of situational awareness is a leading contributor to incidents in dynamic environments such as confined space entry, crane operations, and emergency response.

6. The Dirty Dozen
Originating in aviation maintenance, the Dirty Dozen lists twelve common preconditions for human error: lack of communication, complacency, lack of knowledge, distraction, lack of teamwork, fatigue, lack of resources, pressure, lack of assertiveness, stress, lack of awareness, and norms. The exam may present a scenario and ask which precondition is most evident.

7. Performance Shaping Factors (PSFs)
Used in Human Reliability Analysis (HRA) techniques such as THERP and SPAR-H, PSFs are conditions that increase or decrease the probability of error: time available, stress, complexity, experience and training, procedures, ergonomics and human-machine interface, fitness for duty, and work processes. Know that PSFs multiply a baseline human error probability.

How Decision Making Works and Where It Fails

Two Systems of Thinking
Daniel Kahneman's framework distinguishes System 1 (fast, automatic, intuitive, effortless) from System 2 (slow, deliberate, analytical, effortful). Most daily decisions use System 1, which is efficient but vulnerable to bias. High-consequence decisions should trigger System 2, but fatigue, time pressure, and overconfidence push people back to System 1.

Heuristics and Cognitive Biases Commonly Tested
- Confirmation bias: seeking or weighting information that supports an existing belief (the operator who assumes the alarm is a false alarm because it was last time).
- Availability heuristic: judging likelihood by how easily examples come to mind. Recent or vivid incidents are overweighted; quiet hazards are underweighted.
- Anchoring: over-relying on the first piece of information received.
- Optimism bias and overconfidence: believing bad outcomes are less likely for oneself.
- Sunk cost fallacy and plan continuation bias: continuing a course of action because of prior investment, even when conditions have changed. Classic in weather-related and schedule-pressure incidents.
- Groupthink: suppressing dissent to preserve group harmony, often reinforced by authority gradients.
- Normalization of deviance (Diane Vaughan, from the Challenger investigation): gradual acceptance of deviations from standards because nothing bad happened previously. Each successful deviation lowers the perceived risk until it becomes the new normal.
- Risk homeostasis and risk compensation: people adjust behavior to maintain a target level of perceived risk, so added protection can be offset by riskier behavior.

Naturalistic Decision Making and Recognition-Primed Decision (RPD) Model
Gary Klein showed that experts under time pressure do not compare multiple options; they recognize a pattern, mentally simulate one course of action, and act. This is powerful but fails when the situation is novel or when cues are misleading. Training through realistic scenarios builds the pattern library that RPD relies on.

Stress, Fatigue, and Workload
- The Yerkes-Dodson law describes an inverted-U relationship: moderate arousal optimizes performance, while very low arousal (boredom, complacency) and very high arousal (panic, overload) degrade it.
- Fatigue impairs reaction time, judgment, and vigilance. Being awake 17 to 19 hours produces impairment comparable to a blood alcohol concentration of about 0.05 percent; 24 hours approaches 0.10 percent. Circadian low points occur roughly between 2 and 6 a.m. and again in the mid-afternoon. Fatigue risk management systems (FRMS) address scheduling, shift rotation (forward rotation is preferred), and nap policies.
- Attentional narrowing (tunnel vision) under high stress causes people to fixate on a single cue and miss peripheral information.
- Vigilance decrement: the ability to detect rare signals drops significantly after about 20 to 30 minutes of monitoring, which has implications for inspection tasks and control room design.

Communication Failures
Because this topic sits in the Communication domain, expect questions on: authority gradient (subordinates reluctant to challenge superiors), closed-loop communication (sender states, receiver repeats back, sender confirms), standardized phraseology, shift handover protocols, and assertiveness tools such as graded challenge language. The 1977 Tenerife runway collision is the textbook example of ambiguous phraseology plus authority gradient plus time pressure.

Applying Human Factors: What the Safety Professional Does

Design out the error (Hierarchy of Controls applied to human factors)
Preferred approaches rank roughly as follows: eliminate the error opportunity entirely; use forcing functions and interlocks that physically prevent the wrong action (poka-yoke, mistake-proofing); apply affordances and population stereotypes so controls work the way people expect (clockwise to increase, red for stop, up for on in North America); use shape and color coding and spatial separation to prevent confusion; provide checklists and job aids to offload memory; add independent verification for critical steps; and only then rely on training and warnings. On the exam, if a design fix and a training fix are both offered, the design fix is almost always the better answer.

Design in error tolerance and recovery
Accept that errors will occur and build in detection (alarms, feedback, confirmation prompts) and recovery (undo capability, time delays before irreversible action, redundancy).

Just Culture
A Just Culture distinguishes human error (console and redesign the system), at-risk behavior (coach and remove incentives for shortcuts), and reckless behavior (discipline). It replaces both a blame culture (which drives reporting underground) and a blame-free culture (which fails to address genuine recklessness). The substitution test asks: would a different, similarly qualified person in the same situation likely have done the same thing? If yes, the problem is systemic. Just Culture is essential for the near-miss reporting that feeds learning.

Human Factors in Incident Investigation
Avoid hindsight bias (events look predictable after the fact) and outcome bias (judging a decision by its result rather than by what was known at the time). Ask what made sense to the person at the moment, reconstruct the information they actually had, and identify the PSFs present. Tools include HFACS (Human Factors Analysis and Classification System, which maps Reason's model into unsafe acts, preconditions, unsafe supervision, and organizational influences), TapRooT, and barrier analysis.

Human Factors in Training Design
Training interventions aligned to error type: for slips, use cues, checklists, and reduced distraction rather than more lecture; for lapses, use memory aids and sequence markers; for rule-based mistakes, clarify when rules apply and practice recognizing exceptions; for knowledge-based mistakes, provide mental models, simulation, and access to expertise; for violations, address the underlying incentives, remove conflicting goals, and engage workers in rewriting unworkable procedures. Adult learning principles (relevance, active participation, immediate feedback, spaced repetition) support transfer to the job. Crew Resource Management (CRM) and its industrial analogs train teams in communication, leadership, workload distribution, and challenge-and-response behaviors.

How to Answer Exam Questions on Human Factors and Decision Making

ASP questions in this area are typically scenario-based. A short narrative describes an incident or workplace condition and asks you to identify the error type, the contributing factor, the model that applies, or the best corrective action. Use the following approach:

Step 1: Classify the behavior. Was the action intended? If no, it is a slip or lapse. If the action was intended but the plan was wrong, it is a mistake. If the person knowingly deviated, it is a violation. Getting this classification right unlocks the correct intervention.

Step 2: Identify the performance level. Was the person on autopilot (skill-based), following a procedure (rule-based), or figuring it out (knowledge-based)? This tells you where the system should have supported them.

Step 3: Look for the latent conditions. Scan the scenario for staffing cuts, overtime, ambiguous procedures, similar-looking controls, production pressure, poor lighting, or missing feedback. The exam rewards answers that address these upstream factors.

Step 4: Apply the hierarchy. When multiple fixes are offered, prefer elimination, engineering controls, and forcing functions over procedures, training, discipline, or signage. Reject answers that simply tell workers to pay more attention or that punish a slip.

Step 5: Eliminate distractors. Wrong answers often blame the individual, confuse slips with violations, recommend retraining for what was actually a design problem, or propose discipline where Just Culture would call for coaching.

Worked Example 1
An experienced operator, working the last hour of a 12-hour night shift, closes valve 14B instead of the adjacent, identically shaped valve 14A, releasing product. Which error type and best control?
Analysis: The intention (close 14A) was correct; execution was wrong during a routine task while fatigued. This is a slip at the skill-based level, with fatigue and poor hardware differentiation as PSFs. Best control: physically differentiate the valves (shape, color, position, labeling) and add a verification step, not retraining or discipline.

Worked Example 2
A crew routinely bypasses a machine guard because the production target cannot be met with the guard in place, and supervisors have known this for months. Classification?
Analysis: This is a routine violation with elements of situational violation, sustained by normalization of deviance and conflicting goals. The latent conditions are the unrealistic target and supervisory tolerance. Best response: resolve the production-safety conflict, redesign the task or guard, and address supervision, rather than disciplining only the operators.

Worked Example 3
A new technician, facing an unfamiliar alarm with no procedure, reasons through the situation and makes a wrong diagnosis.
Analysis: Knowledge-based mistake. Error rates are inherently high at this level. Controls: develop a procedure or decision aid, improve alarm design and prioritization, ensure access to expert support, and use simulation-based training.

Exam Tips: Answering Questions on Human Factors and Decision Making

1. Memorize the vocabulary precisely. Slip versus lapse versus mistake versus violation; active failure versus latent condition; skill versus rule versus knowledge-based. Many questions are decided purely on definitions.

2. Default to systems thinking. If an answer choice places responsibility on the system, design, or organization, it is usually more correct than one that places it on the individual. The ASP exam reflects modern safety science, not the old careless-worker paradigm.

3. Training is rarely the best first answer. Training is appropriate for knowledge gaps. It does not reliably fix slips, lapses, fatigue, or design-induced error. If the scenario describes an experienced worker making a routine error, look for a design or environmental fix.

4. Know the numbers that get tested. Approximate error probabilities by SRK level; fatigue equivalence to blood alcohol; the Yerkes-Dodson inverted U; vigilance decrement after about 20 to 30 minutes; circadian low points.

5. Recognize the famous case studies. Challenger (normalization of deviance, groupthink, production pressure), Tenerife (communication, authority gradient), Three Mile Island (interface design, alarm overload, misdiagnosis), Piper Alpha (permit-to-work and shift handover failures), Texas City (fatigue, staffing, instrumentation, organizational pressure). Questions may describe these without naming them.

6. Watch for Just Culture questions. If a worker made an honest error, the correct response is to console and fix the system. If at-risk behavior, coach. Only reckless disregard warrants discipline. An answer that disciplines a slip is wrong.

7. Watch for hindsight bias in the question stem. The exam may present a situation where the right decision seems obvious in retrospect. Evaluate the decision based on what the person knew at the time.

8. For communication questions, favor closed-loop, standardized, and assertiveness-enabling practices. Read-back and verify, standard phraseology, structured handovers, and flattening authority gradients are the recurring correct answers.

9. Link to the hierarchy of controls. Forcing functions, interlocks, and mistake-proofing beat procedures; procedures and checklists beat training alone; training beats warnings and signs.

10. Read every word of the scenario. Details such as end of shift, new employee, identical controls, time pressure, or has happened before are deliberate cues pointing toward fatigue, knowledge-based error, design mismatch, production pressure, or normalization of deviance, respectively.

11. Choose the most proactive, upstream, and sustainable option. When two answers both seem correct, select the one that prevents recurrence across the organization rather than the one that addresses only the immediate event.

12. Manage your own human factors on exam day. Fatigue, time pressure, and anchoring on the first plausible option affect test takers too. Read all four options before selecting, flag uncertain items, and return with a fresh look.

Summary

Human Factors and Decision Making asks the safety professional to understand people as they actually are: limited in attention and memory, shaped by biases and fatigue, responsive to incentives, and operating within systems designed by others. The ASP exam tests whether you can classify errors correctly, identify the latent conditions behind them, apply recognized models such as SHELL, Swiss Cheese, SRK, and Endsley's situational awareness, and select interventions that change the system rather than merely exhort the worker. Master the definitions, think in systems, prefer design over training, and apply Just Culture principles, and you will handle this portion of the exam with confidence.