Safety Culture: 5 Distortions That Make a High Reporting Rate Look Like Trust

A plant receives more safety reports than it did last quarter, and the dashboard turns green. Leaders call the increase proof that trust is improving. The interpretation may be right, but the number alone cannot establish it.

A reporting rate is evidence of information entering the system. It is not a direct measure of psychological safety, control quality, or leadership credibility. The same increase can reflect stronger voice, a campaign that rewards volume, easier digital submission, a new incident classification, or a workforce that has learned that reporting is the only way to obtain action.

This article examines five distortions that make a high reporting rate look like trust. The practical audience is a plant manager or EHS leader who needs to decide whether reporting data is changing the work or merely making activity more visible.

Why a reporting rate is evidence, not a verdict

Safety reporting has value because it brings weak signals, unsafe conditions, near misses, and operational friction into view before the consequences become more severe. James Reason's work on latent failures helps explain why this matters. A visible event is often the last part of a chain that includes design choices, supervision, workload, maintenance, communication, and decisions made far from the task.

The mistake is to treat the number of reports as the outcome. A mature review asks what the reports reveal, how the organization responds, and whether the response changes the conditions that produced the concern. That distinction also reflects Andreza Araújo's position in A Ilusão da Conformidade, known in English as The Illusion of Compliance. The real test of a safety system is what happens when the form, campaign, or observer is no longer present.

In more than 250 cultural transformation projects supported by Andreza Araújo, the useful question is not whether people participated in a ritual. It is whether the organization made safety visible in recurring decisions. A reporting dashboard becomes meaningful only when it helps leaders see and change those decisions.

1. Counting volume without understanding exposure

The first distortion is assuming that more reports automatically mean more trust. Volume can rise because the workforce sees a real hazard, but it can also rise because the definition of reportable activity changed or because a supervisor required a target number from every shift.

Imagine two plants with the same reporting rate. In Plant A, reports describe recurring energy-isolation problems, ambiguous work instructions, and equipment conditions that could cause serious harm. In Plant B, most entries record low-consequence housekeeping observations that are closed with reminders. The dashboard treats the plants as equally active, even though the decision value is different.

Plant leaders should separate report volume from exposure significance. Review the hazard class, credible consequence, task frequency, barrier involved, and recurrence pattern. A high reporting rate with repeated critical conditions may show that the system is finally seeing a problem, not that the problem is being controlled.

The test is simple. For the ten most consequential reports in the period, can the leadership team explain what changed in the work? If the answer is no, the number is measuring detection without demonstrating prevention.

2. Treating every report as equal quality

The second distortion appears when a dashboard rewards the number of submissions without testing the quality of the information. A short entry such as “unsafe behavior observed” is easy to count, yet it gives the next decision-maker almost nothing to verify.

Useful reports describe a condition, an exposure, or a decision barrier. They identify where the work diverged from the intended control and provide enough context for someone else to understand why the concern matters. This does not mean that a worker must write a long narrative. It means the reporting route should make the important facts easier to capture.

Supervisors can improve quality by reviewing whether reports contain a clear condition, the task or location, the potential consequence, the immediate safeguard, and the owner of the next decision. They should also protect the reporter from having the form converted into a performance judgment. When people expect blame, they shorten the description, avoid names, or stop reporting.

A reporting system that produces more entries but weaker evidence may be becoming easier to use while becoming less useful to risk control. Plant leaders should therefore track a small sample of reports from submission to decision, rather than relying on completion counts alone.

3. Confusing submission with response

The third distortion is treating the act of reporting as proof that the organization is listening. A worker may submit a concern because the channel is accessible, because a supervisor instructed the team to use it, or because no other route has produced a response.

Trust is tested after the report arrives. Did someone acknowledge it? Did the organization explain what would happen next? Was the concern routed to a person with authority to change the condition? Did the reporter receive a proportionate explanation without being exposed to retaliation?

Andreza's work on safety culture makes this distinction practical. A culture cannot be installed through a slogan or a reporting application. It is cultivated through time, presence, and consistency, and it is visible in what leaders do with inconvenient information. A report that disappears into an inbox teaches a stronger lesson than a campaign poster can reverse.

Measure response reliability separately from report volume. A useful review can show the share of reports with a named owner, a decision, a communicated outcome, and a later verification. Those fields do not prove that every concern was resolved in the preferred way. They show whether the organization completed the management responsibility created by the report.

4. Measuring activity instead of decision change

The fourth distortion is counting meetings, observations, corrective actions, and closed records as if they were control improvements. Activity matters when it creates a safer condition. It becomes theater when the record is closed while the work remains unchanged.

For each high-consequence report, ask which layer of the work changed. Did the plant alter the design, remove the exposure, strengthen an engineered safeguard, clarify authority, revise the method, improve supervision, or verify the control under normal production pressure? A reminder or retraining session may be appropriate, but it should not be accepted as the default response when the report reveals a design or management failure.

The distinction is consistent with the hierarchy of controls and with the reasoning behind ISO 45001:2018. A management system is not demonstrated by the existence of a procedure. It is demonstrated by the organization's ability to identify hazards, assign responsibility, implement controls, and evaluate whether those controls work.

Plant leaders can improve the dashboard by adding one field that is often missing, namely the decision changed because of the report. The field should describe the operational change, not the meeting held. “Reviewed with the team” records activity. “Installed a physical interlock and changed the authorization boundary” records a control decision.

5. Reading silence as stability

The fifth distortion is the most dangerous. A fall in reporting may look like improvement when it actually reflects fatigue, fear, reporting friction, or the belief that nothing will change.

Silence becomes especially difficult to interpret after a serious concern is dismissed, after a supervisor reacts defensively, or after a reporting campaign ends without visible follow-through. The workforce does not need to announce that trust has fallen. People simply stop investing effort in a channel that appears unable to affect the work.

That is why reporting data needs comparison with other signals. Review field conversations, shift-handover notes, absenteeism patterns, overtime and schedule pressure, repeat defects, maintenance deferrals, and the time between a concern and the management response. None of these signals is a standalone diagnosis. Together, they help leaders avoid treating a quiet system as a safe system.

Use a direct question in leadership reviews: what information would we expect to receive if the workforce saw a serious exposure today? If the answer depends on a single app or a single supervisor, the organization has a channel, not yet a reliable voice system.

What plant leaders should compare before calling reporting healthy

A stronger review compares participation with the quality and consequence of the decisions that follow. The table below gives a practical structure for a monthly plant review.

For a broader evidence model, compare the reporting data with field verification and climate evidence rather than allowing one instrument to define culture. The article on which evidence a plant manager should trust provides a useful companion lens. Worker participation also deserves attention as a decision-quality issue, not only as a count of consultations.

A 30-day review for the plant manager and EHS leader

Start with one reporting stream and one serious exposure category. Do not redesign the entire dashboard before understanding how the current process behaves. Sample recent reports across shifts and departments, then follow each sample from submission to final verification.

• In the first week, classify the reports by exposure, consequence, recurrence, and control layer.
• In the second week, identify where ownership, response, or communication stopped.
• In the third week, select one recurring condition and change the work rather than adding another awareness message.
• In the fourth week, return to the same task during ordinary production and verify whether the control holds.

Keep the review proportionate. The goal is not to turn every report into a formal investigation. The goal is to learn whether the reporting route sends reliable information to people who can make and verify a control decision.

Conclusion: trust is shown by what reports change

A high reporting rate can be encouraging, but it is not self-interpreting. It may reflect stronger voice, higher exposure, better access, a temporary campaign, or a workforce that is trying one more time to obtain a response.

Plant leaders can read the signal more honestly by separating volume from exposure, testing the quality of evidence, measuring response reliability, tracking decision change, and checking whether silence has a credible alternative explanation. That is how participation becomes part of safety culture instead of another activity metric.

Andreza Araújo's books and practical resources develop this connection between declared values and daily decisions. Visit the official Andreza Araújo store for material on safety culture, leadership, and the conditions that make prevention real.