25 Truths About Safety Culture: What the Research Says Leaders Get Wrong

Most safety leaders are working from a rulebook written in 1931. The safety pyramid on the canteen wall, the claim that almost every accident is caused by a careless worker, the bonus for a quarter with no injuries, the "Zero Harm" banner over the gate. All of it feels like common sense. Much of it does not stand up to the evidence.

That matters because the stakes are not abstract. In 2024, 5,070 people died from work injuries in the United States alone, roughly one every 104 minutes (U.S. Bureau of Labor Statistics [BLS], 2026). Worldwide, the International Labour Organization estimates that about 2.93 million workers died from work-related causes in 2019 (International Labour Organization [ILO], 2023). Many of those deaths happened at sites with good-looking injury statistics.

We went through meta-analyses, field experiments, government investigation reports and official guidance to collect the 25 truths about safety culture and safety leadership that the research supports, and that many leaders still get wrong. Each one comes with the evidence, an APA citation you can check, and three practical actions for supervisors, managers and executives in warehouses, transport, logistics and operations.

The short version: Low injury rates do not prove a workplace is safe. Serious injuries and disasters have their own warning signs, and most of them are hidden by the way organisations count, reward and talk about safety. The leaders who make the biggest difference are the frontline supervisors who talk about safety every day, make it safe to report bad news, and look hard at the work that could kill someone.

Key takeaways

• Small injuries fell much faster than deaths. From 1992 to 2019, non-fatal U.S. workplace injuries fell 68.5% while fatal injuries fell only 30%, according to a summary of BLS data by the American Society of Safety Professionals (ASSP, 2021).
• Most injuries never reach the log. On one large construction project, companies with a poor safety climate left 81% of eligible injuries unreported, against 47% at companies with a positive climate (Probst et al., 2008).
• For every accident reported, about 2.5 go unreported. Employees in high-risk industries described an average of 2.48 unreported accidents for every one their organisation heard about (Probst & Estrada, 2010).
• A good injury record did not save Texas City. The 2005 BP refinery explosion killed 15 workers and injured 180 at a site where, investigators found, safety was judged mainly by counting personal injuries (U.S. Chemical Safety and Hazard Investigation Board [CSB], 2007).
• Supervisors rarely talk about safety. Before coaching, construction workers said safety came up in only 6–16% of their conversations with foremen. Coaching foremen to talk about it raised site safety levels by 7% and 12% (Kines et al., 2010).
• Leadership style changes injury rates. Safety-specific transformational leadership predicted fewer injuries in two separate studies of 174 and 164 workers (Barling et al., 2002).

How to read this guide: The 25 truths are grouped into four parts. Citations in brackets lead to the full reference list at the bottom. This article is general information for leaders. It is not legal, engineering or medical advice. For legal duties, follow the regulator and the law where you operate, and use competent safety professionals for hazard-specific controls.

Part 1: The myths leaders inherited

Truth 1: The "88% of accidents are caused by unsafe acts" figure was never proven

In 1931, insurance engineer H. W. Heinrich claimed that 88% of industrial accidents were caused by the unsafe acts of workers, 10% by unsafe conditions and 2% were unpreventable (Heinrich, 1931). The figure has been repeated in safety training for nearly a century. Safety researcher Fred Manuele reviewed Heinrich's work and concluded that the premise is not soundly based and should be dislodged from safety practice (Manuele, 2011). Heinrich's original data have never been made available for anyone to check, and the classification relied on how supervisors described accidents after the fact. Modern safety science points the other way: unsafe acts usually reflect the design of the work, the tools, the time pressure and the organisation around the person (Reason, 1997; Hofmann et al., 2017).

What leaders should do:

• Stop quoting the 88% figure in inductions, toolbox talks and board papers.
• When you read "worker failed to follow procedure" in a report, ask what made that shortcut normal, likely or necessary.
• Put at least as much budget into fixing conditions and equipment as into behaviour campaigns.

Truth 2: Fewer minor injuries does not mean fewer deaths

Heinrich also gave us the triangle: for every 330 accidents of the same kind, he suggested, 300 cause no injury, 29 cause minor injuries and 1 causes a major injury (Heinrich, 1931). Many leaders took the lesson that driving down small injuries would automatically shrink the top of the pyramid. Manuele (2011) called this the second myth. The national numbers agree with him. From 1992 to 2019, non-fatal U.S. workplace injuries fell 68.5%, but fatal injuries fell only 30% (ASSP, 2021). A literature review found many industries now stuck on a plateau, with low minor-injury rates and surprise fatal accidents that bear no relation to how they were measuring safety (Dekker & Pitzer, 2016).

What leaders should do:

• Track serious injuries, fatalities and high-potential events as a separate line on every safety report.
• Don't celebrate a fall in total recordables until you know whether serious events moved too.
• Ask your safety team which of your tasks could kill someone, and check what controls protect each one.

Truth 3: Serious injuries and fatalities have their own warning signs

If deaths don't simply scale with minor injuries, what predicts them? Research on serious injuries and fatalities (SIF) argues that they have distinct "precursors": high-risk situations such as work at height, energy isolation, mobile equipment and lifting, where controls are missing or weak (Martin & Black, 2015). The National Safety Council's Campbell Institute recommends that employers identify these precursors, train people to recognise them and redesign work so it relies less on individual workers getting everything right (Campbell Institute, 2018). One industry estimate, from a safety consultancy quoted by the National Safety Council's magazine, is that about 21% of OSHA-recordable events had realistic potential to cause a serious injury or fatality (Safety+Health, 2017). That figure is a practitioner estimate, not a peer-reviewed rate, but it shows why sorting incidents by potential, not just actual outcome, matters.

What leaders should do:

• Classify every incident and near miss by its potential severity, not only by what happened.
• Investigate high-potential near misses as seriously as you would investigate an actual serious injury.
• List your site's top SIF exposures, such as pedestrian and forklift interaction, loading docks and working at height, and audit their controls in person.

Truth 4: "Human error" is where an investigation should start, not where it ends

Psychologist James Reason distinguished two ways of seeing error. The "person approach" blames individuals for forgetfulness or carelessness. The "system approach" accepts that people make mistakes and asks why the defences failed (Reason, 2000). Reason argued that the person approach is attractive because it is emotionally satisfying and legally convenient, but it isolates unsafe acts from their context and stops learning. Sidney Dekker calls the alternative the "new view" of human error: error is a symptom of trouble deeper inside the system, so investigations should explain why the person's actions made sense to them at the time (Dekker, 2014b).

What leaders should do:

• Ban "human error" and "lack of attention" as final root causes in your investigation forms.
• Ask investigators to describe the situation as the worker saw it: time pressure, information available, tools and goals.
• Track how many corrective actions change equipment, layout or process, not just retraining or discipline.

Truth 5: A site can look safe on paper right up until it explodes

On 23 March 2005, a tower at BP's Texas City refinery overfilled during a startup. The release ignited, killing 15 workers and injuring 180 (CSB, 2007). The CSB found long-standing process safety deficiencies and cost-cutting that left the refinery vulnerable, while BP judged its safety performance mainly by counting injuries to individuals (CSB, 2007). An independent panel led by former U.S. Secretary of State James Baker reached a similar conclusion: BP had treated improving personal injury rates as a sign that its process safety was acceptable, and needed to give process safety the same priority as personal safety (BP U.S. Refineries Independent Safety Review Panel, 2007). Andrew Hopkins argued that personal injury rates and process safety indicators measure different things, and that one cannot stand in for the other (Hopkins, 2008, 2009).

What leaders should do:

• If your operation has major hazards, such as fuel, chemicals, pressure or heavy moving equipment, report process safety indicators separately from slips, trips and falls.
• Treat overdue inspections, maintenance backlogs and repeated alarms as safety data, not just engineering data.
• When cost cuts touch maintenance or staffing on safety-critical work, require a written risk assessment before approval.

Truth 6: Senior leaders can walk right past the hazard that matters

On 20 April 2010, four senior managers from BP and Transocean were aboard the Deepwater Horizon on a management visit. Hours later the Macondo well blew out, and 11 workers died (CSB, 2016a). Hopkins (2011) analysed the visit and found that, despite good intentions, the visitors focused on personal safety topics and paid almost no attention to the safety-critical well operations happening around them. The CSB's final report concluded that a culture of minimal regulatory compliance persisted offshore, and called for leading process safety indicators and a process safety culture improvement programme (CSB, 2016a, 2016b). The presidential commission likewise traced the disaster to systemic management failures, not one bad decision (National Commission on the BP Deepwater Horizon Oil Spill and Offshore Drilling, 2011).

What leaders should do:

• Before a site visit, ask: "What is the most dangerous work happening here today?" and go and look at it.
• Use visits to ask how major hazards are being controlled, not only to check PPE and housekeeping.
• Ask frontline staff what worries them most, and report back on what you did about it.

Part 2: What actually works: climate, supervisors and leadership

Truth 7: Safety climate predicts injuries, especially at team level

Safety climate is what employees believe is really valued and rewarded when safety competes with speed or cost (Zohar, 1980). A meta-analysis by Christian et al. (2009) found that safety knowledge and motivation were most strongly linked to safe behaviour, but when it came to actual accidents and injuries, group safety climate had the strongest association. Zohar (2000) showed this in 53 work groups in one manufacturing company: groups with better climate perceptions had fewer minor injuries over the following five months. Clarke's (2006) meta-analysis added a caution: the link between climate and accidents held up reliably only in prospective studies, where climate was measured before the accidents.

What leaders should do:

• Measure safety climate by team or shift, not just as one site-wide score.
• Look for the gaps between teams. The weakest team's climate is your real risk.
• Use a validated climate survey and repeat it, so you can see change over time.

Truth 8: Injuries shape climate more than climate shapes injuries

Beus et al. (2010) separated the two directions in a meta-analysis. They found that injuries were more predictive of later organisational safety climate than climate was of later injuries. In plain terms, people's view of safety partly reflects the accidents they have already seen. The climate-to-injury link also weakened as the time period lengthened. Neal and Griffin (2006), tracking accidents over five years, found that group climate predicted later safety motivation, which predicted safer behaviour, and that improved group behaviour was followed by fewer accidents. Climate works, but slowly and through people.

What leaders should do:

• Don't treat one good survey result as proof of a safe site. Read it alongside incident and near-miss data.
• After a serious incident, expect climate to dip, and show visible action fast.
• Plan for change over years, not quarters. Climate shifts follow sustained behaviour from leaders.

Truth 9: Workers judge management commitment by what happens under pressure

Of all the dimensions of safety climate, perceived management commitment to safety was the most robust predictor of occupational injuries in the Beus et al. (2010) meta-analysis. Huang et al. (2006) found that management commitment, return-to-work policies, post-injury administration and safety training formed core parts of climate, and that climate predicted self-reported injury through employees' sense of control over their safety. Employees don't learn commitment from posters. They learn it from what happens when a job is late, a truck is waiting or a machine needs fixing mid-shift (Zohar, 2002).

What leaders should do:

• Publicly back a worker who stops a job for safety, even when it costs time or money.
• Make sure people who are injured are treated fairly and helped back to work. Everyone is watching how that goes.
• Never let a production target be met by quietly skipping a safety step.

Truth 10: Frontline supervisors are the real safety system

Hofmann and Stetzer (1996) found that group processes and safety climate were linked to unsafe behaviours and accidents across work teams. Zohar then tested whether changing supervisors' behaviour could change outcomes. In a field experiment, line supervisors received weekly feedback on how often they had safety-oriented conversations with their teams. Safety-oriented interaction rose in the experimental groups, and minor-injury rates, earplug use and safety climate scores improved and stayed improved (Zohar, 2002). Three follow-up interventions in other companies produced the same pattern of better safety behaviour and climate (Zohar & Luria, 2003).

What leaders should do:

• Make safety conversations a measured part of every supervisor's job, not an add-on.
• Give supervisors regular feedback on how often and how well they talk about safety.
• Promote supervisors partly on how safely their teams work, not only on output.

Truth 11: Safety-specific transformational leadership reduces injuries

Barling et al. (2002) studied 174 restaurant workers, then replicated the model with 164 young workers from many jobs. Leaders who showed safety-specific transformational leadership, meaning they inspired, set an example, cared about individuals and challenged people to think about safety, had teams with better safety climate and safety consciousness, and that predicted fewer injuries. A later meta-analysis found that transformational leadership was linked to safety participation, while active transactional leadership (monitoring and correcting) was linked to both participation and compliance (Clarke, 2013). The two styles work best together.

What leaders should do:

• Explain the "why" behind safety rules with real stories, not just the rule itself.
• Ask your team for their ideas on making the work safer, and act on the good ones visibly.
• Combine inspiration with follow-through: check, recognise and correct safety behaviour consistently.

Truth 12: Passive safety leadership is not neutral. It does harm

Many managers believe that if they leave safety to the safety department, they are at worst neutral. Kelloway et al. (2006) showed otherwise. Safety-specific passive leadership, such as avoiding decisions, waiting for problems to become serious and not responding to concerns, was a distinct pattern from transformational leadership. It added to the prediction of safety outcomes beyond transformational leadership alone, and it had the opposite effect: it weakened safety climate and consciousness, which in turn predicted more safety events and injuries.

What leaders should do:

• Respond to every safety concern within an agreed time, even if the answer is "not yet, and here's why".
• Don't wait for an incident to address a hazard you already know about.
• Ask your team to rate how quickly you act on safety issues, and treat a poor score as a performance issue.

Truth 13: Safety leadership can be trained, and talk is the lever

If leadership matters, can it be taught? Mullen and Kelloway (2009) ran a longitudinal study with health care managers comparing general leadership training, safety-specific leadership training and a control group, and linked the training to safety outcomes among employees. In Danish construction, Kines et al. (2010) coached foremen and gave them bi-weekly feedback on their daily safety talk. At baseline, workers said safety featured in only 6–16% of their exchanges with foremen. After coaching, safety communication rose by a factor of 4.6 across the intervention groups, and observed site safety levels improved by 7% and 12%, while control groups did not change. Visibility helps too: supervisors who could easily see their workers kept up safety interactions because they saw the results (Luria et al., 2008).

What leaders should do:

• Coach supervisors to include one specific safety point in their daily conversations with each crew.
• Give supervisors short, regular feedback on their safety talk, not a one-off course.
• Design layouts and routines so supervisors are physically present where the risky work happens.

Truth 14: Engagement and support beat pure compliance

Wachter and Yorio (2014) studied ten safety management practices and found that both the practices and workers' safety-focused cognitive and emotional engagement were linked to lower accident rates. Engagement acted as a bridge: management systems worked partly by winning workers' minds and hearts. A meta-analysis of 203 samples and 186,440 people found that job demands such as hazards and complexity drove burnout, which was linked to less safe working, while job resources, especially a supportive environment, drove engagement and safer working in every industry studied (Nahrgang et al., 2011).

What leaders should do:

• Involve workers in writing procedures, risk assessments and investigations for their own tasks. OSHA lists worker participation as a core element of any safety programme (OSHA, 2016b).
• Watch for fatigue and overload. Staffing and scheduling are safety decisions.
• Build support in: quick access to help, working equipment and time to do the job properly.

Part 3: The measurement trap: reporting, incentives and zero

Truth 15: Most injuries never make it into the official numbers

Probst et al. (2008) compared contractors' OSHA logs with insurance medical claims on a large construction site. The reported injury rate was 3.11 per 100 workers, but the rate of eligible injuries not reported to OSHA was 10.90 per 100 workers. Companies with poor safety climate failed to report 81% of eligible injuries, against 47% for companies with positive climate. At the national level, Rosenman et al. (2006) matched Michigan data sources and estimated that the national surveillance system missed 61%, and up to 68%, of work-related injuries and illnesses. The cost is huge either way: occupational injury and illness cost the U.S. an estimated $250 billion in 2007 (Leigh, 2011).

What leaders should do:

• Treat a sudden drop in reported injuries as a question, not a win, until you know why.
• Compare your injury logs with first-aid records, clinic visits and insurance claims to spot gaps.
• Count near-miss and hazard reports as a sign of a healthy culture, and expect them to rise when trust grows.

Truth 16: Production pressure quietly switches off reporting

Probst and Graso (2013) surveyed 212 copper miners. On average, each worker had experienced 2.84 accidents but reported only 0.49. Workers who felt high production pressure had more negative attitudes to reporting, experienced more accidents and reported fewer of them. Across five high-risk industries, Probst and Estrada (2010) found 2.48 unreported accidents for every one reported, with under-reporting worse where climate was poor or supervisors enforced safety inconsistently. When leaders push output hard, they don't just raise risk. They also blind themselves to it.

What leaders should do:

• Say clearly and often that reporting an injury or near miss will never count against a person or team.
• During peak periods, increase your safety walk-throughs rather than cutting them.
• Ask workers anonymously whether they feel pressure to keep quiet about minor injuries.

Truth 17: Injury-free bonuses can buy silence instead of safety

Pransky et al. (1999) surveyed 110 workers at three industrial facilities. Fewer than 5% had officially reported a work-related injury or illness in the past year, yet more than 85% had work-related symptoms. Workers cited fear of reprisal, and managers described barriers rooted in a goal of zero reported injuries. The authors judged that safety incentives had an indirect but significant negative influence on reporting. The U.S. Government Accountability Office estimated that about a quarter of U.S. manufacturers had safety incentive programmes, with 22% using rate-based programmes that reward low injury numbers (U.S. Government Accountability Office [GAO], 2012). In 2012, OSHA warned that prize draws or team bonuses for going injury-free can discourage reporting (Occupational Safety and Health Administration [OSHA], 2012).

What leaders should do:

• Never tie bonuses, prizes or league tables to having no reported injuries.
• Reward the inputs instead: hazards fixed, near misses reported, safety ideas implemented.
• Review any team target that a single honest injury report could wipe out.

Truth 18: Retaliating against reporters is illegal, and OSHA can act without a complaint

In 2016, OSHA amended its recordkeeping rule to require that employers tell workers they have the right to report injuries without retaliation, that reporting procedures be reasonable, and that employers not retaliate against workers who report (OSHA, 2016a). OSHA later clarified that the rule does not ban safety incentive programmes or post-incident drug testing outright. They violate the rule only when they are used to penalise people for reporting (OSHA, 2018). The legal detail differs by country, but the principle is universal: if your system punishes people for telling the truth, you will stop hearing it. This is not legal advice; check your own jurisdiction.

What leaders should do:

• Tell every worker, in writing and in person, how to report an injury and that retaliation is prohibited.
• Review discipline and drug-testing policies with legal advice so they never punish the act of reporting.
• Audit your reporting procedure: if it takes several forms and a manager's signature, it is probably discouraging reports.

Truth 19: "Zero harm" targets can backfire

Zero is a powerful word, and a moral aspiration that nobody should be hurt is reasonable (Zwetsloot et al., 2013). The trouble starts when zero becomes a numerical target. Sherratt and Dainty (2017) examined four years of UK Health and Safety Executive data alongside major contractors' safety approaches. Working on a project with a zero policy appeared to slightly increase the likelihood of a serious, life-changing accident or fatality. They urged caution with that finding, but warned that zero policies can stifle learning without reducing accidents. Dekker et al. (2016) argued that zero vision can become a belief system in which every incident is a moral failure, which pushes bad news underground.

What leaders should do:

• If you use "zero", frame it as a commitment that no one should be seriously hurt, not as a reporting target.
• Never present an incident as a personal failure of the team that reported it.
• Pair any zero message with targets for learning, such as near misses investigated and fixes completed.

Truth 20: Lagging indicators measure nothing when things look good

Injury rates and lost-time counts are lagging indicators: they count harm after it happens. When incidents are rare, these numbers stop telling you much. Lofquist (2010) described this as "measuring nothing" in aviation, where traditional metrics failed to show how organisational change was affecting safety. OSHA's guidance encourages leading indicators, which show whether safety activities are working before people are hurt (OSHA, 2019). A meta-analysis of construction studies identified leading indicators that correlate with injury rates (Alruqi & Hallowell, 2019), and other researchers have shown how leading measures can monitor and steer an organisation's capacity to stay safe (Hinze et al., 2013; Reiman & Pietikäinen, 2012).

What leaders should do:

• Pick three to five leading indicators that fit your risks, such as critical-control checks completed, corrective actions closed on time and near-miss reports.
• Check the quality behind the numbers. A checklist ticked without looking is a lagging indicator in disguise.
• Put leading indicators at the top of the monthly report and injury rates underneath.

Part 4: The new view: drift, deviance and what goes right

Truth 21: The most dangerous risks are the ones you have stopped noticing

Sociologist Diane Vaughan studied the 1986 Challenger launch decision and described the "normalisation of deviance": small departures from expected performance were accepted, nothing bad happened, and the departures became the new normal (Vaughan, 1996). The Columbia Accident Investigation Board later found that foam strikes had followed the same path, so that flying with known flaws became routine before Columbia was lost in 2003 (Columbia Accident Investigation Board [CAIB], 2003). Disaster researchers describe a similar "incubation period" in which warning signs build up unnoticed before a man-made disaster (Pidgeon & O'Leary, 2000).

What leaders should do:

• Ask each team: "What do we do here that would shock a newcomer?" Then look at the answers seriously.
• Treat any workaround that has become routine as a hazard to review, not a sign of good teamwork.
• Bring in outside eyes, such as people from another site, to walk your operation and challenge what you accept.

Truth 22: Systems drift toward danger one sensible decision at a time

Jens Rasmussen showed that organisations are pushed by pressure for efficiency and by people's natural wish to reduce effort. Together these move work gradually toward the boundary of safe performance, without anyone deciding to take a big risk (Rasmussen, 1997). Each step looks reasonable in the moment. Dekker and Pitzer (2016) linked this to the plateau in safety performance, arguing that compliance-heavy practices can create a sense of invulnerability, divert resources into unproductive activity and encourage the suppression of bad news.

What leaders should do:

• When you approve a change in staffing, speed, shift pattern or equipment, ask what it does to safety margins.
• Review the combined effect of small changes once a year, not just each change on its own.
• Reward people who flag that margins are getting thin, even when nothing has gone wrong yet.

Truth 23: Safety is also about why things usually go right

Erik Hollnagel calls the traditional approach "Safety-I": safety means as few things as possible going wrong, so you study failures and remove their causes. "Safety-II" defines safety as the ability to succeed under varying conditions, so you also study everyday work to understand why it usually goes right (Hollnagel, 2014). Hollnagel et al. (2015) argued that the two approaches are complementary, not rival. Provan et al. (2020) described what this means for safety professionals: supporting guided adaptation by frontline workers, not only enforcing central control.

What leaders should do:

• Spend some safety walk time asking "How do you get this done safely on a bad day?" rather than only looking for faults.
• Study high-performing shifts and crews to learn what they do differently.
• Give experienced workers room to adapt to real conditions, within clear limits on critical controls.

Truth 24: Work as imagined is not work as done, and more paperwork won't close the gap

Hale and Borys (2013) reviewed decades of research on safety rules. They contrasted a top-down view, in which rules written by experts must be followed, with a bottom-up view, in which rules emerge from the expertise of the people doing the work. Both have a place, but rules that ignore real work get bent, quietly. Dekker (2014a) argued that safety has become increasingly bureaucratised, with growing paperwork and compliance activity that does not always make work safer and can crowd out the practical expertise that does.

What leaders should do:

• Ask the people who do the job to review each critical procedure: "What does this get wrong about how the work really happens?"
• Remove one low-value form or sign-off for every new one you add.
• Keep firm, non-negotiable rules for life-critical controls, and keep them few and clear.

Truth 25: Bad news only travels upward where it is safe to send

Amy Edmondson's study of 51 work teams found that team psychological safety, a shared belief that it is safe to take interpersonal risks, was linked to learning behaviour, and learning behaviour drove performance (Edmondson, 1999). In safety, the same mechanism decides whether near misses and mistakes are reported. Dekker and Breakey (2016) argued that a just culture means more than deciding who to blame: it should restore trust, repair harm and help the people involved learn. As the reporting studies show, under-reporting is worst where climate is poor and supervision inconsistent (Probst & Estrada, 2010).

What leaders should do:

• Thank people every time they report a mistake or near miss, especially their own.
• After an incident, ask "What happened and how do we help?" before "Who did this?"
• Share what changed because of a report, so people see that speaking up works.

Your 10-point safety leadership action plan

The 25 truths point to the same handful of moves. None of them needs a big budget. All of them need consistent leadership.

1. Retire the myths. Drop the 88% figure and the idea that fewer small injuries automatically prevent deaths.
2. Separate serious from minor. Track serious injuries, fatalities and high-potential events on their own line.
3. Hunt for SIF precursors. Know your deadliest tasks and check their critical controls in person.
4. Make supervisors the safety system. Measure, coach and reward daily safety conversations.
5. Show commitment under pressure. Back people who stop work for safety, especially when it costs you.
6. Fix incentives. Never reward low injury counts. Reward reporting, fixes and safety ideas.
7. Protect reporting. Make it easy, make retaliation impossible and treat rising near-miss reports as good news.
8. Lead with leading indicators. Put critical-control checks and corrective actions above injury rates.
9. Watch for drift. Ask what has quietly become normal, and review the combined effect of small changes.
10. Learn from work as done. Study what goes right, involve workers in procedures and cut paperwork that doesn't make work safer.

Ready to lead safety, not just report it? Explore BLTI's logistics, transportation and operations credentials, built for supervisors and managers who run real-world operations, and take the free Leadership Flex Check to see how you lead when pressure is on.

Frequently asked questions

What is the difference between safety culture and safety climate?

Safety culture usually refers to the deeper, shared values and assumptions about safety in an organisation. Safety climate is the measurable snapshot of what employees believe is really valued and rewarded when safety competes with other goals (Zohar, 1980). Climate is what most surveys measure, and group-level climate has the strongest link to accidents and injuries (Christian et al., 2009).

Is Heinrich's safety triangle still valid?

Not in the way it is often taught. Heinrich's original data have never been verified, and critics argue that reducing minor incidents does not proportionally reduce serious injuries (Manuele, 2011). U.S. data show non-fatal injuries falling much faster than fatal ones (ASSP, 2021). Many organisations now focus on serious injury and fatality precursors instead.

Are safety incentive programmes legal?

In the U.S., OSHA's position is that incentive programmes are not banned, but they must not be used to penalise workers for reporting injuries (OSHA, 2016a, 2018). Programmes that reward low injury counts can discourage reporting, so many experts recommend rewarding safe activities instead (Pransky et al., 1999). Laws differ by country, so take legal advice for your jurisdiction.

What are leading indicators in safety?

Leading indicators measure the activities and conditions that prevent harm, such as hazard reports, completed inspections, critical-control verifications and corrective actions closed on time. Lagging indicators, such as injury rates, count harm after it happens. OSHA recommends using both, with leading indicators showing whether your programme is working before anyone is hurt (OSHA, 2019).

What is Safety-II?

Safety-II, developed by Erik Hollnagel, defines safety as the ability to succeed under varying conditions. Instead of only studying what goes wrong, it also studies everyday work to understand why things usually go right (Hollnagel, 2014). It complements traditional safety management rather than replacing it (Hollnagel et al., 2015).

Is "zero harm" a good safety goal?

As a commitment that nobody should be seriously hurt, it can be useful (Zwetsloot et al., 2013). As a numerical target, the evidence suggests it can discourage reporting and learning, and one UK study found zero-policy projects did not have fewer serious accidents (Sherratt & Dainty, 2017).

About this research

This guide was compiled by the Better Leaders Training Institute (BLTI) research team. Every source below was checked for authors, year, title, publisher and DOI or stable link, using publisher records, Crossref, PubMed and official agency websites. We prioritised peer-reviewed studies, meta-analyses, official government data and investigation reports from bodies such as the U.S. Chemical Safety Board, OSHA, the Bureau of Labor Statistics and the ILO. Where a figure comes from a practitioner estimate, a professional association's summary or trade press, we say so in the text. Findings from individual studies describe specific samples and settings; your operation may differ. This article is general information, not legal, engineering or medical advice.

References

Alruqi, W. M., & Hallowell, M. R. (2019). Critical success factors for construction safety: Review and meta-analysis of safety leading indicators. Journal of Construction Engineering and Management, 145(3), Article 04019005. https://doi.org/10.1061/(ASCE)CO.1943-7862.0001626

American Society of Safety Professionals. (2021, May 24). Three keys to preventing serious injuries and fatalities. https://www.assp.org/news/three-keys-to-preventing-serious-injuries-and-fatalities

Barling, J., Loughlin, C., & Kelloway, E. K. (2002). Development and test of a model linking safety-specific transformational leadership and occupational safety. Journal of Applied Psychology, 87(3), 488–496. https://doi.org/10.1037/0021-9010.87.3.488

Beus, J. M., Payne, S. C., Bergman, M. E., & Arthur, W., Jr. (2010). Safety climate and injuries: An examination of theoretical and empirical relationships. Journal of Applied Psychology, 95(4), 713–727. https://doi.org/10.1037/a0019164

BP U.S. Refineries Independent Safety Review Panel. (2007). The report of the BP U.S. Refineries Independent Safety Review Panel. BP.

Campbell Institute. (2018). Serious injury and fatality prevention: Perspectives and practices. National Safety Council. https://www.thecampbellinstitute.org/

Christian, M. S., Bradley, J. C., Wallace, J. C., & Burke, M. J. (2009). Workplace safety: A meta-analysis of the roles of person and situation factors. Journal of Applied Psychology, 94(5), 1103–1127. https://doi.org/10.1037/a0016172

Clarke, S. (2006). The relationship between safety climate and safety performance: A meta-analytic review. Journal of Occupational Health Psychology, 11(4), 315–327. https://doi.org/10.1037/1076-8998.11.4.315

Clarke, S. (2013). Safety leadership: A meta-analytic review of transformational and transactional leadership styles as antecedents of safety behaviours. Journal of Occupational and Organizational Psychology, 86(1), 22–49. https://doi.org/10.1111/j.2044-8325.2012.02064.x

Columbia Accident Investigation Board. (2003). Report: Volume I. National Aeronautics and Space Administration.

Dekker, S. W. A. (2014a). The bureaucratization of safety. Safety Science, 70, 348–357. https://doi.org/10.1016/j.ssci.2014.07.015

Dekker, S. (2014b). The field guide to understanding 'human error' (3rd ed.). Ashgate.

Dekker, S. W. A., & Breakey, H. (2016). 'Just culture:' Improving safety by achieving substantive, procedural and restorative justice. Safety Science, 85, 187–193. https://doi.org/10.1016/j.ssci.2016.01.018

Dekker, S. W. A., Long, R., & Wybo, J.-L. (2016). Zero vision and a Western salvation narrative. Safety Science, 88, 219–223. https://doi.org/10.1016/j.ssci.2015.11.016

Dekker, S., & Pitzer, C. (2016). Examining the asymptote in safety progress: A literature review. International Journal of Occupational Safety and Ergonomics, 22(1), 57–65. https://doi.org/10.1080/10803548.2015.1112104

Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350–383. https://doi.org/10.2307/2666999

Hale, A., & Borys, D. (2013). Working to rule, or working safely? Part 1: A state of the art review. Safety Science, 55, 207–221. https://doi.org/10.1016/j.ssci.2012.05.011

Heinrich, H. W. (1931). Industrial accident prevention: A scientific approach. McGraw-Hill.

Hinze, J., Thurman, S., & Wehle, A. (2013). Leading indicators of construction safety performance. Safety Science, 51(1), 23–28. https://doi.org/10.1016/j.ssci.2012.05.016

Hofmann, D. A., Burke, M. J., & Zohar, D. (2017). 100 years of occupational safety research: From basic protections and work analysis to a multilevel view of workplace safety and risk. Journal of Applied Psychology, 102(3), 375–388. https://doi.org/10.1037/apl0000114

Hofmann, D. A., & Stetzer, A. (1996). A cross-level investigation of factors influencing unsafe behaviors and accidents. Personnel Psychology, 49(2), 307–339. https://doi.org/10.1111/j.1744-6570.1996.tb01802.x

Hollnagel, E. (2014). Safety-I and Safety-II: The past and future of safety management. Ashgate.

Hollnagel, E., Wears, R. L., & Braithwaite, J. (2015). From Safety-I to Safety-II: A white paper. The Resilient Health Care Net. https://www.england.nhs.uk/signuptosafety/wp-content/uploads/sites/16/2015/10/safety-1-safety-2-whte-papr.pdf

Hopkins, A. (2008). Failure to learn: The BP Texas City refinery disaster. CCH Australia.

Hopkins, A. (2009). Thinking about process safety indicators. Safety Science, 47(4), 460–465. https://doi.org/10.1016/j.ssci.2007.12.006

Hopkins, A. (2011). Management walk-arounds: Lessons from the Gulf of Mexico oil well blowout. Safety Science, 49(10), 1421–1425. https://doi.org/10.1016/j.ssci.2011.06.002

Huang, Y.-H., Ho, M., Smith, G. S., & Chen, P. Y. (2006). Safety climate and self-reported injury: Assessing the mediating role of employee safety control. Accident Analysis & Prevention, 38(3), 425–433. https://doi.org/10.1016/j.aap.2005.07.002

International Labour Organization. (2023). A call for safer and healthier working environments. ILO. https://www.ilo.org/topics-and-sectors/safety-and-health-work

Kelloway, E. K., Mullen, J., & Francis, L. (2006). Divergent effects of transformational and passive leadership on employee safety. Journal of Occupational Health Psychology, 11(1), 76–86. https://doi.org/10.1037/1076-8998.11.1.76

Kines, P., Andersen, L. P. S., Spangenberg, S., Mikkelsen, K. L., Dyreborg, J., & Zohar, D. (2010). Improving construction site safety through leader-based verbal safety communication. Journal of Safety Research, 41(5), 399–406. https://doi.org/10.1016/j.jsr.2010.06.005

Leigh, J. P. (2011). Economic burden of occupational injury and illness in the United States. The Milbank Quarterly, 89(4), 728–772. https://doi.org/10.1111/j.1468-0009.2011.00648.x

Lofquist, E. A. (2010). The art of measuring nothing: The paradox of measuring safety in a changing civil aviation industry using traditional safety metrics. Safety Science, 48(10), 1520–1529. https://doi.org/10.1016/j.ssci.2010.05.006

Luria, G., Zohar, D., & Erev, I. (2008). The effect of workers' visibility on effectiveness of intervention programs: Supervisory-based safety interventions. Journal of Safety Research, 39(3), 273–280. https://doi.org/10.1016/j.jsr.2007.12.003

Manuele, F. A. (2011). Reviewing Heinrich: Dislodging two myths from the practice of safety. Professional Safety, 56(10), 52–61.

Martin, D. K., & Black, A. A. (2015). Preventing serious injuries and fatalities. Professional Safety, 60(9).

Mullen, J. E., & Kelloway, E. K. (2009). Safety leadership: A longitudinal study of the effects of transformational leadership on safety outcomes. Journal of Occupational and Organizational Psychology, 82(2), 253–272. https://doi.org/10.1348/096317908X325313

Nahrgang, J. D., Morgeson, F. P., & Hofmann, D. A. (2011). Safety at work: A meta-analytic investigation of the link between job demands, job resources, burnout, engagement, and safety outcomes. Journal of Applied Psychology, 96(1), 71–94. https://doi.org/10.1037/a0021484

National Commission on the BP Deepwater Horizon Oil Spill and Offshore Drilling. (2011). Deep water: The Gulf oil disaster and the future of offshore drilling. Report to the President. U.S. Government Printing Office. https://www.govinfo.gov/app/details/GPO-OILCOMMISSION

Neal, A., & Griffin, M. A. (2006). A study of the lagged relationships among safety climate, safety motivation, safety behavior, and accidents at the individual and group levels. Journal of Applied Psychology, 91(4), 946–953. https://doi.org/10.1037/0021-9010.91.4.946

Occupational Safety and Health Administration. (2012, March 12). Employer safety incentive and disincentive policies and practices [Memorandum]. U.S. Department of Labor. https://www.osha.gov/laws-regs/standardinterpretations/2012-03-12-0

Occupational Safety and Health Administration. (2016a). Improve tracking of workplace injuries and illnesses. Federal Register, 81(92), 29624–29694. https://www.osha.gov/laws-regs/federalregister/2016-05-12

Occupational Safety and Health Administration. (2016b). Recommended practices for safety and health programs (OSHA 3885). U.S. Department of Labor. https://www.osha.gov/safety-management

Occupational Safety and Health Administration. (2018, October 11). Clarification of OSHA's position on workplace safety incentive programs and post-incident drug testing under 29 C.F.R. § 1904.35(b)(1)(iv) [Memorandum]. U.S. Department of Labor. https://www.osha.gov/laws-regs/standardinterpretations/2018-10-11

Occupational Safety and Health Administration. (2019). Using leading indicators to improve safety and health outcomes (OSHA 3970). U.S. Department of Labor. https://www.osha.gov/leading-indicators

Pidgeon, N., & O'Leary, M. (2000). Man-made disasters: Why technology and organizations (sometimes) fail. Safety Science, 34(1–3), 15–30. https://doi.org/10.1016/S0925-7535(00)00004-7

Pransky, G., Snyder, T., Dembe, A., & Himmelstein, J. (1999). Under-reporting of work-related disorders in the workplace: A case study and review of the literature. Ergonomics, 42(1), 171–182. https://doi.org/10.1080/001401399185874

Probst, T. M., Brubaker, T. L., & Barsotti, A. (2008). Organizational injury rate underreporting: The moderating effect of organizational safety climate. Journal of Applied Psychology, 93(5), 1147–1154. https://doi.org/10.1037/0021-9010.93.5.1147

Probst, T. M., & Estrada, A. X. (2010). Accident under-reporting among employees: Testing the moderating influence of psychological safety climate and supervisor enforcement of safety practices. Accident Analysis & Prevention, 42(5), 1438–1444. https://doi.org/10.1016/j.aap.2009.06.027

Probst, T. M., & Graso, M. (2013). Pressure to produce = pressure to reduce accident reporting? Accident Analysis & Prevention, 59, 580–587. https://doi.org/10.1016/j.aap.2013.07.020

Provan, D. J., Woods, D. D., Dekker, S. W. A., & Rae, A. J. (2020). Safety II professionals: How resilience engineering can transform safety practice. Reliability Engineering & System Safety, 195, Article 106740. https://doi.org/10.1016/j.ress.2019.106740

Rasmussen, J. (1997). Risk management in a dynamic society: A modelling problem. Safety Science, 27(2–3), 183–213. https://doi.org/10.1016/S0925-7535(97)00052-0

Reason, J. (1997). Managing the risks of organizational accidents. Ashgate.

Reason, J. (2000). Human error: Models and management. BMJ, 320(7237), 768–770. https://doi.org/10.1136/bmj.320.7237.768

Reiman, T., & Pietikäinen, E. (2012). Leading indicators of system safety: Monitoring and driving the organizational safety potential. Safety Science, 50(10), 1993–2000. https://doi.org/10.1016/j.ssci.2011.07.015

Rosenman, K. D., Kalush, A., Reilly, M. J., Gardiner, J. C., Reeves, M., & Luo, Z. (2006). How much work-related injury and illness is missed by the current national surveillance system? Journal of Occupational and Environmental Medicine, 48(4), 357–365. https://doi.org/10.1097/01.jom.0000205864.81970.63

Safety+Health. (2017, August 27). Preventing serious injuries and fatalities. National Safety Council. https://www.safetyandhealthmagazine.com/?p=25521

Sherratt, F., & Dainty, A. R. J. (2017). UK construction safety: A zero paradox? Policy and Practice in Health and Safety, 15(2), 108–116. https://doi.org/10.1080/14773996.2017.1305040

U.S. Bureau of Labor Statistics. (2026, February 19). National census of fatal occupational injuries in 2024 [News release]. U.S. Department of Labor. https://www.bls.gov/news.release/cfoi.nr0.htm

U.S. Chemical Safety and Hazard Investigation Board. (2007). Investigation report: Refinery explosion and fire, BP Texas City, Texas, March 23, 2005 (Report No. 2005-04-I-TX). https://www.csb.gov/bp-america-texas-city-refinery-explosion/

U.S. Chemical Safety and Hazard Investigation Board. (2016a). Investigation report volume 3: Drilling rig explosion and fire at the Macondo well (Report No. 2010-10-I-OS). https://www.csb.gov/macondo-blowout-and-explosion/

U.S. Chemical Safety and Hazard Investigation Board. (2016b, April 13). The U.S. Chemical Safety Board's investigation into the Macondo disaster finds offshore risk management and regulatory oversight still inadequate in Gulf of Mexico [News release]. https://www.csb.gov/the-us-chemical-safety-boards-investigation-into-the-macondo-disaster-finds-offshore-risk-management-and-regulatory-oversight-still-inadequate-in-gulf-of-mexico/

U.S. Government Accountability Office. (2012). Workplace safety and health: Better OSHA guidance needed on safety incentive programs (GAO-12-329). https://www.gao.gov/products/gao-12-329

Vaughan, D. (1996). The Challenger launch decision: Risky technology, culture, and deviance at NASA. University of Chicago Press.

Wachter, J. K., & Yorio, P. L. (2014). A system of safety management practices and worker engagement for reducing and preventing accidents: An empirical and theoretical investigation. Accident Analysis & Prevention, 68, 117–130. https://doi.org/10.1016/j.aap.2013.07.029

Zohar, D. (1980). Safety climate in industrial organizations: Theoretical and applied implications. Journal of Applied Psychology, 65(1), 96–102. https://doi.org/10.1037/0021-9010.65.1.96

Zohar, D. (2000). A group-level model of safety climate: Testing the effect of group climate on microaccidents in manufacturing jobs. Journal of Applied Psychology, 85(4), 587–596. https://doi.org/10.1037/0021-9010.85.4.587

Zohar, D. (2002). Modifying supervisory practices to improve subunit safety: A leadership-based intervention model. Journal of Applied Psychology, 87(1), 156–163. https://doi.org/10.1037/0021-9010.87.1.156

Zohar, D., & Luria, G. (2003). The use of supervisory practices as leverage to improve safety behavior: A cross-level intervention model. Journal of Safety Research, 34(5), 567–577. https://doi.org/10.1016/j.jsr.2003.05.006

Zwetsloot, G. I. J. M., Aaltonen, M., Wybo, J.-L., Saari, J., Kines, P., & Op De Beeck, R. (2013). The case for research into the zero accident vision. Safety Science, 58, 41–48. https://doi.org/10.1016/j.ssci.2013.01.026

Videos

Edmondson, A. (2014, May 4). Building a psychologically safe workplace [Video]. TEDx Talks. https://www.youtube.com/watch?v=LhoLuui9gX8

U.S. Chemical Safety and Hazard Investigation Board. (2008, June 2). CSB safety video: Anatomy of a disaster [Video]. YouTube. https://www.youtube.com/watch?v=XuJtdQOU_Z4

U.S. Chemical Safety and Hazard Investigation Board. (2014, June 5). Deepwater Horizon blowout animation [Video]. YouTube. https://www.youtube.com/watch?v=FCVCOWejlag