5 Insights from Episode 11 with Dr. Thomas Krause on Leadership Quality

Episode 11 of Headline Podcast was published on December 3, 2025, with Dr. Thomas Krause, CEO of the Krause Bell Group, in conversation with Andreza Araujo and Dr. Megan Tranter. Krause's central argument was that safety success depends less on declared values than on the quality of leadership people experience in daily decisions.

Why does leadership quality predict safety success?

Leadership quality predicts safety success because workers experience safety through decisions, not through the wording of a value statement. A leader who asks about production first, accepts a weak control, or closes a concern without returning to the field teaches the organization what really matters, regardless of what the policy says.

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.” The point is not that leaders must perform constant visibility. It is that their choices need to make the intended standard easier to follow when time, cost, and uncertainty compete.

Andreza Araujo's work across more than 250 cultural transformation projects reaches the same practical question. Does the system make the safe decision more credible, more supported, and more repeatable than the shortcut? If the answer is no, a communication campaign will not repair the leadership signal.

OSHA's Recommended Practices for Safety and Health Programs describe management leadership and worker participation as connected elements, while ISO 45001:2018 specifies leadership and worker participation requirements. Both sources point toward an operating system in which leaders remain accountable for conditions, resources, and follow-through.

Insight 1: Value becomes credible when workers can feel it

A safety value becomes credible when workers can predict what a leader will do after inconvenient evidence appears. Posters and speeches can introduce an aspiration, but only a repeated response to risk tells the workforce whether the aspiration has operational force.

Krause described this distinction as value that is felt rather than declared. A supervisor who pauses a job, protects the person who raised the concern, and returns with a decision creates a stronger cultural signal than a leader who repeats “safety first” during a quarterly meeting.

The test can be run in 30 days. Select 3 recent concerns, record the original response, identify the decision owner, and ask the affected crew whether the final action changed the work. The answer should include evidence, not only satisfaction. A closed conversation is not the same as a changed exposure.

In *Safety Culture: From Theory to Practice*, Andreza Araujo frames culture through repeated choices under pressure. That framing helps leaders avoid a common trap, which is to measure belief before checking whether the system supports the behavior that belief requires.

Insight 2: Culture lives between people, not on posters

Culture lives between people because the most consequential safety decisions are interpreted in conversations between supervisors, operators, engineers, and managers. The formal procedure may be identical across 2 shifts, while the practical standard changes according to who has authority and how dissent is received.

A leader can examine this gap by comparing 4 moments. What happens when a worker challenges a plan? What happens when a permit is incomplete? What happens when a target is missed because a control was restored? What happens when an incident has no obvious individual error?

Those moments reveal whether respect is conditional. If the person who raises a concern is praised in public but excluded from the next planning meeting, the organization has learned that speaking up carries a cost. The stated culture remains positive while the lived culture becomes cautious.

NIOSH recommends examining organizational and work conditions that shape safety and health, which gives leaders a stronger lens than asking whether employees have the right attitude. The relevant unit is the interaction, the decision, and the consequence that follows.

Insight 3: Incident analysis must look beyond the last visible action

Incident analysis becomes useful when it explains why the work made the final action likely, difficult, or invisible. Stopping at the person who touched the equipment produces a simple story, but it leaves the design, planning, supervision, and resource decisions that shaped the event untouched.

Dr. Thomas Krause put the problem directly: “Incident analysis always looks like it was the employee's fault for not following the procedure, but look deeper and you see following it was made very difficult by system factors set by decisions made a year, or five years, ago.”

The practical response is to build an evidence chain across at least 5 layers. Review the task condition, the procedure, the equipment or design, the supervisor's decision, and the earlier management choice that established the constraint. Then separate facts from assumptions before assigning action.

Within the first 24 hours, preserve records, identify witnesses, and protect the scene from unnecessary change. Within 7 days, test whether the proposed correction changes the work rather than only reminding people to be careful. James Reason's work on latent failures supports this distinction because visible error can be the final expression of older system conditions.

Insight 4: Trust scores need a denominator leaders can see

A trust score can reassure a leader while concealing a large minority that no longer expects a fair response. The useful question is not whether the average looks positive. It is whether the score identifies who is silent, where trust breaks down, and which management decision follows.

Krause offered a sharp example on the episode. “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 don't trust their boss.” The number is not a case for rejecting surveys. It is a warning against treating rank as proof of health.

Use 3 cuts when reviewing a trust measure. Separate shifts, roles, and tenure groups; compare the score with reported concerns and stop-work events; then ask whether low-scoring groups receive a visible response. A score without segmentation tells leaders how the average feels, not where the operating risk is concentrated.

Andreza Araujo's experience in more than 30 countries reinforces the need for local interpretation. The same percentage can describe different leadership problems in a refinery, a distribution center, and a corporate office, so the decision must follow the work context rather than the ranking alone.

Insight 5: The floor conversation is a control check

A field conversation is a control check when it tests what people understand, what they can do, and what they will do when the plan stops matching reality. A walk that only confirms housekeeping or collects compliments creates visibility without verification.

Ask 5 questions in the work area. What can hurt someone here? Which control prevents it? What would show that the control has failed? Who can stop the job? What will happen after you raise the concern? The answers should be specific enough to connect a person, a barrier, and a decision.

The leader should then verify one answer in the field. If the crew names an isolation, inspect the isolation. If they name an alarm, test the alarm or review the latest functional evidence. If they name stop-work authority, ask how the last pause was handled and whether the work plan changed afterward.

This routine can take 10 minutes, yet its value depends on the return loop. A leader who promises to check and never comes back teaches people that the conversation is ceremonial. A leader who returns within 48 hours with evidence closes the loop and strengthens the next conversation.

What changes when leadership quality is the operating variable?

Leadership quality changes the diagnosis from “employees need more commitment” to “the system needs more credible decisions.” That shift matters because the same workforce can produce different safety results when leaders change priorities, authority, resources, and responses to bad news.

The table is useful because it separates language from proof. A leader does not need to create 20 new initiatives. The stronger move is to select 3 recurring decisions and make their evidence, ownership, and follow-up visible.

Recommendation

Start with one leadership-quality review over the next 30 days. Choose 3 recent safety concerns, 1 incident analysis, and 1 field conversation, then compare what leaders declared with what workers experienced.

For each item, name the decision owner, the control being tested, the evidence that supports the current status, and the date for returning to the people affected. If the answer is missing, do not hide the gap inside a new score. Escalate the decision and make the uncertainty visible.

Use OSHA's guidance, ISO 45001:2018, and NIOSH's organizational-safety resources as reference points, but keep the diagnosis local. A standard can define expectations; only leadership behavior can show whether the organization makes those expectations real.

Andreza Araujo has spent more than 25 years in multinational safety leadership and has documented how culture changes when decision rights, worker voice, and operational follow-through meet. Read the wider work on Andreza Araujo's safety leadership resources, then Listen to the full conversation with Dr. Thomas Krause.

A positive survey result can coexist with silent risk. The next leadership review should test the 40% that the average score may leave unseen.