Incident Evidence: 8 Myths That Make Investigations Less Reliable

An incident investigation becomes less reliable when a complete-looking record is treated as proof that the event has been understood. A photograph can show a condition without showing its history. A witness can describe a sequence without remembering every interruption. A procedure can define intended work while the crew operates under different constraints.

Investigators must separate observation from interpretation, compare independent sources, and keep uncertainty visible until the evidence supports a conclusion. The goal is to identify which conditions allowed the exposure and which changes will prevent a recurrence.

Why evidence quality matters more than report volume

A long report can still be weak when it contains repeated descriptions, unsupported assumptions, and actions that do not change the work. James Reason’s work on organizational accidents is useful because it distinguishes active failures at the workface from latent conditions embedded in planning, design, resources, and supervision.

The investigator’s job is not to collect everything. It is to collect enough reliable evidence to test competing explanations and show where the operating system made the exposure possible.

Myth 1: The first account is the most accurate account

The first account is often the most available account, not the most accurate one. People reconstruct a sequence from partial views, sounds, expectations, and conversations after the event.

Interview promptly, but do not turn speed into certainty. Ask what the witness saw, heard, touched, and decided before asking why the decision made sense. Compare the account with equipment status, permits, radio traffic, and the physical scene. “The operator saw the valve open” is different from “the valve was opened by the operator.” The first may be an observation. The second requires corroboration.

Myth 2: A photograph preserves the whole scene

A photograph preserves a frame, not the event. It can show a guard, spill, tool, or damaged component, but it rarely shows what changed before the camera arrived. The angle may hide access, lighting, line of sight, stored energy, or the sequence of work.

Photograph broadly before moving to detail, then record location, time, orientation, lighting, and changes made for rescue or stabilization. When the area has changed, mark the image as a reconstruction and describe its basis.

Myth 3: The procedure is what happened

A procedure shows the intended method. It does not prove that the method was available, understood, supervised, or workable under the conditions present that day.

Read the procedure alongside the permit, task plan, equipment condition, staffing, shift handover, and local instructions. Ask which step was difficult to execute and what the team did when the written sequence met physical reality. This does not excuse a departure. It makes the departure analyzable.

Myth 4: A witness statement is a factual transcript

A witness statement is a person’s account of an experience. It can contain direct observations, assumptions, later explanations, and language borrowed from other conversations.

Use open prompts before testing a theory. Ask what happened next, what the person expected, and what information was available at the decision point. When accounts conflict, record the conflict instead of averaging them into a convenient sentence. A neutral process improves evidence quality when the interview concerns production pressure or known equipment defects.

Myth 5: The last action explains the event

The last action is visible, which makes it attractive as an explanation. Yet it usually occurs inside a chain of preparation, information, equipment, supervision, and competing priorities.

Test what the person knew, what the person could see, which controls were available, and what would have happened if the person had stopped. Then trace earlier decisions that affected those conditions. A behavior can be part of the causal chain without being the complete cause.

Myth 6: A signed training record proves competence

A training record proves that an administrative event was recorded. It does not prove that the person understood the hazard, could perform the task, or had the equipment and time required to apply the instruction.

Review practical demonstration, supervision, task complexity, language, changing conditions, and the opportunity to ask questions. Training is weak when it is used to avoid redesigning equipment, improving isolation, clarifying ownership, or removing conflicting production demands.

Myth 7: A root cause is a single sentence

A single sentence can summarize an explanation, but serious incidents often involve interacting conditions. The phrase “operator error” may describe the final action while hiding a poor interface, ambiguous instruction, delayed maintenance response, and supervision gap.

Use a causal map or timeline to show how conditions connected. Andreza Araújo’s book A Ilusão da Conformidade, translated as The Illusion of Compliance, is relevant because a documented rule can create a false sense of control when operating conditions have not been verified. A root-cause sentence should point toward a condition that can be changed and checked.

Myth 8: Closing the action closes the risk

An action is not complete because a person uploaded a document, signed a training list, or marked a task closed. Closure means the exposure changed and the new control works under the conditions that produced the event.

Define acceptance criteria before assigning the action. The owner must have authority to make the change, while the verifier must be able to challenge a weak result. If the work still depends on memory, heroics, or silent workarounds, the risk has not been closed.

How to build a more defensible evidence trail

Write a timeline with three columns. Record what was observed, what source supports it, and what remains uncertain. Then compare the evidence across the physical scene, the human account, the work system, and barrier performance.

When the evidence does not support a confident conclusion, state the uncertainty and identify the next verification step. Honest limits are stronger than false precision when leaders decide whether similar exposure exists elsewhere.

For a practical review routine, compare this evidence trail with the five decision gaps that turn a timeline into a story and the evidence-preservation steps to take before the worksite changes.

What investigators should ask before issuing the report

Ask whether each important statement is an observation, an inference, or an unanswered question. Ask whether the investigation tested the work as performed, not only as the procedure describes it. Ask whether the proposed actions change the conditions that made the exposure possible.

If the answer is only that people were reminded to be careful, the investigation has not reached a reliable control. Safety Culture: From Theory to Practice connects culture with routines, decisions, and consequences. The report should make the next safe decision easier for the person doing the work and more visible to the leader responsible for the system.

Make investigations change the work. Explore Andreza Araújo’s safety culture resources for methods that connect evidence, leadership, and prevention.

Frequently Asked Questions