How Dr. Thomas Krause Thinks About Making Safety Leadership Visible

Episode 11 of the Headline Podcast, published on December 3, 2025, asked why some safety initiatives look active while the leadership behind them remains invisible.

Dr. Thomas Krause defended a demanding answer: safety improves when leadership quality is experienced in daily decisions, not when a value is merely printed, announced, or measured once a year.

Leadership quality is the first safety variable to examine

Leadership quality is the strongest practical test of whether a safety initiative can survive pressure, because leaders allocate attention, resources, credibility, and decision rights when the work becomes difficult. In Episode 11, Dr. Thomas Krause said that the strongest predictor of success was the quality of leadership given to the initiative.

On Headline Podcast, Dr. Thomas Krause said, "We were surprised to learn the strongest predictor of success was the quality of leadership given to the initiative."

That observation changes the starting point for an EHS review. Instead of asking only whether the program has a policy, a steering group, or a training calendar, ask what leaders do when production, maintenance, cost, and safety appear to compete. The answer is usually visible in minutes, approvals, staffing choices, and the questions asked after a deviation.

OSHA reports that management leadership includes the vision and resources needed to implement an effective safety and health program. The standard is useful because it treats leadership as operating behavior rather than executive messaging.

A practical test uses 3 recent decisions. Did a leader stop work when a control became uncertain? Did the leader protect time for worker participation? Did the leader fund a corrective action that had no immediate production return? Those decisions reveal more than a poster or a slogan.

Declared safety values fail when workers cannot feel them

A safety value becomes credible when workers can predict how it will affect a real decision within 24 hours, such as a staffing change, a permit delay, or a request for help. A declared value that never changes those decisions remains a communication artifact rather than a cultural condition.

Krause's argument is especially important for organizations that publish ambitious commitments. The question is not whether people have heard the message. The question is whether the message changes what happens when a supervisor reports a weak barrier at 4:30 p.m., when a contractor asks for clarification, or when a team says the procedure cannot be followed under current conditions.

Andreza Araujo's Safety Culture: From Theory to Practice makes a related distinction between declared culture and the conditions people actually experience. On Headline, that idea is best used as a test for credibility. If the value is real, workers should be able to point to a decision that proves it.

Use 4 evidence points during a leadership review. Check the last stop-work decision, the last overdue corrective action, the last worker concern escalated, and the last plan changed because field conditions differed from the assumption. Those records make an abstract value observable.

Culture lives between people, not on posters

Safety culture is created in repeated interactions, so a poster cannot compensate for a supervisor who punishes bad news or a manager who rewards speed after a control has weakened. The relevant unit is the conversation between people who plan, supervise, perform, maintain, and authorize the work.

This is why a floor visit should not be reduced to checking whether signs are present. A leader needs to hear how a worker describes uncertainty, how a supervisor responds to challenge, and whether the team can name the person who may change the plan. The quality of those exchanges is an operating signal.

OSHA recommends worker participation in developing and implementing every element of a safety and health program, while also protecting workers from retaliation for reporting hazards. That requirement gives leaders a concrete benchmark. Participation is not a meeting count. It is evidence that worker information changes the system.

Choose 2 work areas and ask the same 5 questions: What changed today? Which control do you trust most? What would make it less reliable? Who can pause the task? What happens after you raise a concern? Compare the answers with management's assumptions rather than grading workers on confidence.

Incident analysis should look beyond the last visible action

An incident review becomes useful when it traces how earlier decisions shaped the conditions in which the final action occurred. The last visible action matters, but it may be only the final link in a chain that began 1 year or 5 years earlier through design, staffing, procurement, planning, or leadership choices.

Krause warned that incident analysis often makes the event look like an employee's failure to follow a procedure, even though following it was made difficult by system factors established much earlier. That warning is not an excuse to ignore individual accountability. It is a demand to place the action inside the work system that made it likely.

James Reason's work on latent failures provides a defensible way to apply this principle without weakening responsibility. Investigators can examine the active failure, the missing barrier, the decision context, and the management condition that allowed the exposure to remain possible.

Before approving an investigation, ask for 3 layers of evidence. First, what did the person do? Second, what made the safe action difficult, unclear, or unrecoverable? Third, which leader-owned condition allowed that difficulty to persist? The third question is where leadership becomes part of prevention.

Trust metrics can hide the people leadership is losing

Averages can reassure leaders while concealing a large minority that does not trust its supervisor. Krause used a sharp example from the episode: 60% of people may say they trust their supervisor while the result still places the organization in the 90th percentile, leaving 40% without that trust.

On Headline Podcast, Dr. Thomas Krause said, "You could have 60% of people say they trust their supervisor and still be in the 90th percentile, so the leader pats himself on the back while 40% of his people do not trust their boss."

The implication is not that every survey score is invalid. It is that a single average cannot tell a leader where silence is forming. Teams, shifts, contractors, locations, tenure groups, and reporting lines can experience the same company very differently.

Separate the headline score into at least 4 views, then compare them with operational evidence. A high trust result paired with low concern reporting may indicate that people are being polite, uncertain about consequences, or unclear about how escalation works.

Internal guidance on psychological safety can help here. Compare the trust result with the number of concerns raised, the median time to response, the percentage of actions closed on time, and the number of repeated concerns. The relationship between those measures is more informative than the score alone.

Controls support leadership only when people can challenge them

A control supports leadership when people can explain its purpose, identify its limits, and challenge it before conditions make it unreliable. NIOSH describes a hierarchy in which elimination, substitution, and engineering controls generally reduce exposure with less dependence on human interaction, but even strong controls require verification when work changes.

Leaders should therefore ask 6 control questions. What hazard is the control intended to address? Which assumption must remain true? How will a worker know the control is degraded? Who can stop the task? What response is available? What evidence shows that the control still works?

ISO 45001:2018 specifies requirements for an occupational health and safety management system, and ISO describes the system as a way to manage risks and improve performance. A certificate, however, does not answer the field questions by itself. Leadership must connect the system to decisions made during ordinary work.

Review 3 critical controls in the next 30 days. Observe them under normal conditions, test one credible deviation, and record whether a worker can escalate without waiting for an incident. That exercise turns leadership from a value statement into a control-verification routine.

Recommendation

Safety leaders should make leadership visible through a small number of repeatable decisions, because visibility gives workers evidence about what the organization will protect under pressure. Start with one operational area, one critical control, one recent concern, and one investigation that has already been closed.

For a practical leadership cadence, compare these checks with the five safety-leadership decisions that turn a weekly plan into visible care.

After 30 days, publish what changed, what did not change, and which decision still depends on senior approval. That last item matters because a culture of visible care is not built by claiming that the organization is finished. It is built by showing where responsibility sits and what leaders will do next.

For a related governance perspective, read Headline's comparison of safety committees, safety councils, and line leadership. For the worker voice dimension, see the Headline companion on neurodiversity as a psychological safety test.

Listen to the full Episode 11 conversation with Dr. Thomas Krause to hear the leadership argument in his own words.