Witness Statements: 7 Interview Errors After Incidents

Witness statements can make an ICAM incident investigation stronger, but they can also make it weaker when the first interview only records the theory that leaders already believe. This article shows seven interview errors that turn memory into noise and explains how EHS managers can protect facts before RCA begins.

The thesis is direct: a witness statement is not a conversation about who failed. It is an evidence control whose quality depends on timing, separation, neutral wording, and connection with the physical work system.

Why witness statements fail before RCA starts

Witness statements fail when organizations treat interviews as administrative paperwork rather than evidence preservation. A person who saw the event may remember the sequence clearly at first, but memory changes when coworkers discuss the incident, when supervisors suggest a cause, or when leadership asks for a fast explanation.

As Andreza Araujo argues in Safety Culture: From Theory to Practice, culture is revealed by what people consider normal under pressure. That matters after an incident because the witness is not only describing an event. The witness is describing the work culture that made certain decisions, shortcuts, controls, and silences feel acceptable at the time.

This article is for EHS managers and supervisors who already have incident evidence preservation routines but still see weak interviews feeding weak RCA. The goal is to make each statement useful enough to test controls, sequence, supervision, and decision pressure without turning the witness into the accused.

1. Waiting until witnesses have aligned their stories

The first interview error is delay. A delayed statement is not automatically false, but it becomes harder to separate original memory from repeated conversations, phone messages, hallway theories, and what the person later learned from others. In high-potential events, the first individual account should be captured as soon as the person is safe enough to speak.

Across 25+ years leading EHS at multinationals, Andreza Araujo has observed that the earliest stories often contain the weak signals that later disappear from formal reports. 25+ years of multinational EHS experience is not a substitute for evidence, but it reinforces a practical rule: time changes what people emphasize.

The fix is simple in design and demanding in execution. Separate witnesses respectfully, explain that the first account is not a disciplinary interview, and ask for a short written or recorded description before anyone presents a cause. If the person needs medical care, emotional support, or union representation, protect those rights while still preventing group contamination of facts.

This timing discipline links directly with post-incident action planning because actions based on contaminated statements often solve the wrong problem. A fast corrective action can look decisive while leaving the real control weakness untouched.

2. Asking leading questions that confirm the first theory

A leading question tells the witness what answer the interviewer expects. Questions such as why did the operator bypass the guard, why was the rule ignored, or who approved the shortcut already frame the event as individual failure before the investigation has tested design, supervision, workload, and control status.

James Reason's work on human error helps here because it separates visible actions from latent conditions. If the interviewer only asks about the visible act, the statement will rarely reveal the conditions that made the act likely. The better question is what made this task look normal, available, necessary, or acceptable at that moment.

Use neutral prompts that cover perception and context: what did you see, what did you hear, what task was planned, what changed, what instruction was given, what control was expected, what made the work proceed, and what seemed unusual. The interviewer should not argue, correct, complete sentences, or reward the answer that matches the early theory.

3. Treating witness protection as softness

Witness protection is an evidence requirement, not a kindness added after the technical work. People who fear blame, retaliation, ridicule, or job loss edit their statements, especially when the event involves a serious injury, a contractor, a supervisor decision, or a production target that everyone understood but nobody documented.

In more than 250 cultural transformation projects supported by Andreza Araujo's team, interview quality improves when workers believe the organization wants the operating truth rather than a convenient culprit. 250+ cultural transformation projects show a repeated pattern: fear narrows testimony, while disciplined respect expands it.

Protection does not mean the investigation avoids accountability. It means the interviewer separates accountability from accusation. The witness can describe an unsafe act, a poor decision, or a rule breach while the team still asks which control, signal, constraint, habit, or leadership routine made that path plausible.

This is also where post-incident meetings can either protect voice or destroy it. A public meeting that hints at blame before interviews finish teaches witnesses to become careful, vague, and less useful.

4. Recording statements without testing them against evidence

A witness statement is evidence, but it is not the whole evidence package. Memory must be compared with photographs, CCTV, access logs, alarms, permits, JSA documents, LOTO records, maintenance history, tool condition, training records, and supervisor walk notes. Without that comparison, the statement becomes an opinion with a signature.

Andreza Araujo's Portuguese title A Ilusao da Conformidade, translated as The Illusion of Compliance, is useful because many investigations confuse documented compliance with real control. A witness may say the permit was reviewed, while the field evidence shows that the control listed on the permit was not present, not understood, or not usable.

Build a statement-to-evidence matrix for serious events. For each claim, mark whether it is confirmed, contradicted, unresolved, or outside the scope of available evidence. This protects the witness from unfair certainty and protects the organization from building RCA around a memory that the scene does not support.

5. Ignoring what the witness believed was normal

The most important sentence in a witness interview may not be what happened. It may be we always do it this way, everyone knows that guard sticks, the lift plan is copied from the last job, or the supervisor was nearby and did not stop it. Normalization of deviance often enters the investigation through casual language.

During the PepsiCo South America tenure, where the accident ratio fell 50% in six months under a 180-day plan, Andreza Araujo learned that repeated normal work patterns explain more than isolated decisions. 50% accident-ratio reduction in six months came from changing routines, leadership presence, and control discipline, not from better slogans.

The interviewer should therefore ask what usually happens on this task, what was different this time, what people normally check, which shortcut is tolerated, and which condition has been reported before. These questions turn a witness statement into a culture sample, which is far more useful than a narrow account of the last visible movement.

6. Letting hierarchy enter the interview room

Hierarchy distorts witness statements when the person being interviewed knows that a supervisor, manager, contractor representative, or senior leader is listening. Even silent authority changes what people say, especially in operations where speaking against production pressure has consequences.

The better practice is to define who may attend each interview before the incident occurs. EHS, HR, legal, union, and management roles should be clear, but attendance should not intimidate the witness or create conflicts of interest. A supervisor whose decision is part of the timeline should not sit inside a subordinate's first account.

For SIF potential events, create a protected interview route with documented rights, neutral location, and clear confidentiality boundaries. The organization should also record who was present, why they were present, and whether the witness requested support. This is not bureaucracy. It is how the team protects the credibility of the statement.

7. Turning the interview into a confession hunt

The final error is treating the interview as a confession hunt. Once the interviewer searches for admission, witnesses begin defending themselves, protecting friends, or guessing what the organization wants to hear. The investigation loses the operating truth that would help prevent recurrence.

Five Whys for SIFs becomes weak when the first why is aimed at a person instead of the work system. The stronger route asks why the control did not prevent exposure, why the work plan made the exposure possible, why supervision did not detect drift, and why previous weak signals did not trigger action.

The interviewer can still ask difficult questions. The difference is posture. Ask for sequence, context, decisions, signals, constraints, and control expectations. Then test the answers against evidence. A statement that admits a rule breach may be important, although it is incomplete until the team understands why the breach was possible, normal, or rewarded.

Witness statements versus weak interview habits

Every day that witness interviews remain informal increases the chance that your next RCA will sound precise while resting on aligned memories, missing context, and leadership assumptions.

Conclusion

Witness statements protect incident facts when they are collected early, separated from influence, shaped by neutral questions, and tested against control evidence. They damage the investigation when they become paperwork, confession hunting, or a way to confirm the first theory spoken by the most senior person in the room.

If your organization needs to improve incident interviews, RCA inputs, and safety culture diagnosis, work with Andreza Araujo to redesign the investigation routine. Safety is about coming home, and people come home more often when investigations protect the truth early enough to change the system.