Everybody trains the leaders. who develops everyone else?

Psychosocial safety can not be built through leadership development alone.

The leadership development day is in the boardroom. Fourteen people, catered lunch, an external facilitator and a workbook with the organisation's values printed inside the front cover. It's a good session. The facilitator is experienced, the material is sound, and the leaders in the room will probably leave with a better understanding of communication, mental health and psychological safety.

Four hundred metres away, three hundred people are working a shift. They're making decisions, responding to pressure, navigating conflict, solving problems and influencing one another in hundreds of small ways.

During that shift, someone may notice that a colleague isn't themselves. Someone may see a task that doesn't look safe, hear a comment that crosses a line or make a mistake and decide whether to admit it. Someone may disagree with a supervisor and quietly calculate whether speaking up is worth the risk.

Almost every organisation that begins taking psychosocial safety seriously arrives at the same response: develop the leaders.

It's a defensible instinct. Leaders influence workload, role clarity, communication, trust and the behaviour that's rewarded or tolerated. Research involving more than 2,800 physicians found that higher supervisor leadership scores were associated with lower odds of burnout and greater job satisfaction.¹ A meta-analysis covering more than 22,000 workers found that positive leader relationships and supportive work environments were among the strongest correlates of psychological safety.² A systematic review of workplace bullying also found that passive and authoritarian leadership were repeatedly associated with greater risk of harm.³

Leaders matter, that's not in dispute.

But psychosocial safety is created through the interaction of systems, leadership and human capability.

Organisations have invested heavily in the first two, or at least claimed to, while largely overlooking the everyday people whose state, perception, communication and behaviour ultimately determine what happens in the moment.

What leadership training can and can't do

In 2022, the World Health Organization published its first global guidelines on mental health at work. The recommendations were built from 36 systematic reviews, additional primary studies and an international guideline development process.⁴

The findings on manager mental health training are revealing.

Manager training produced strong improvements in managers' knowledge. It had a smaller effect on stigmatising attitudes and a very small effect on the mental health of the people they supervised.⁴

That doesn't mean manager training is worthless.

A manager who understands mental health, recognises warning signs and knows how to respond supportively is better equipped than one who doesn't. Knowledge matters, reduced stigma matters, and the behaviour of a supervisor can change the experience of an entire team.

But the evidence suggests that developing the manager doesn't automatically change the conditions or capabilities of everyone working beneath them.

The manager learns, and they may even think differently. But the rest of the workforce still returns to the same workload, the same roster, the same relationships, the same pressures and the same personal patterns they brought to work that morning.

At its best, leadership development is a necessary part of change. At its worst, it becomes a proxy for change. The organisation can point to the workshop, the attendance register and the leadership framework while the lived experience of work remains largely untouched.

Psychosocial safety has three parts

Psychosocial safety cannot be built through leader development alone. It requires three things working together:

• Organisational conditions that reduce harm

• Leaders who model and reinforce safe behaviour

• A workforce equipped with the human capabilities required to communicate, regulate, think and act under pressure

When one of those pieces is missing, the responsibility placed on the other two becomes unrealistic.

1. The conditions of work

The strongest evidence we currently have supports changing the work itself.

A systematic review of 22 controlled studies involving more than 6,300 healthcare workers found that organisational interventions improved at least one mental health or wellbeing outcome in 15 of the 22 studies. Burnout improved in 11 of the 13 studies that measured it, with the strongest evidence found for changes to jobs and tasks.⁵

These changes included the practical conditions people work within, such as workload, scheduling, workflow, task design, professional supports and the physical work environment.

The Job Demands-Resources framework helps explain why this matters. Every job contains demands that consume physical, emotional or cognitive energy, as well as resources that help people meet those demands, recover and remain engaged.⁶

Job demands	Job resources
High workload	Supportive supervision
Fatigue and insufficient recovery	Clear expectations
Role conflict or ambiguity	Adequate staffing
Emotional labour	Autonomy and participation
Poor communication	Useful feedback
Unpredictable or unsafe rostering	Skill development
Procedural friction	Practical tools and support
Low control over work	Meaningful recovery time

When demands repeatedly outweigh the resources available, people don't simply need a better attitude. Their capacity is being depleted by the design of the work.

No amount of capability building can make an unsafe roster safe. A communication workshop can't compensate for chronic understaffing, and a resilience program can't fix a culture where speaking honestly leads to punishment.

The conditions matter because they shape what people are being asked to cope with in the first place.

2. The behaviour leaders make normal

The second piece is leadership.

Leaders shape much more than strategy. They influence what can be said, what gets hidden, whose concerns are taken seriously and whether mistakes become opportunities for learning or reasons for blame.

Amy Edmondson's foundational research established psychological safety as a shared team belief that people can take interpersonal risks without being punished or humiliated.⁷ Teams with stronger psychological safety were more willing to ask for help, admit mistakes, test ideas and learn in public.

Later research has reinforced the importance of both leader behaviour and the wider work environment. Psychological safety isn't simply a personality trait held by confident people. It's shaped by what happens when someone takes a risk.²

What happens when a worker asks a question or challenges a decision? What happens when a new employee admits they don't understand, or when someone says they're overwhelmed?

The formal policy might say one thing, but the leader's response teaches the real lesson.

Leadership training is valuable when it changes that response, but training alone isn't accountability. A leader may understand psychological safety and still silence people under pressure. They may know the correct language and still reward overwork. They may complete mental health training and still dismiss the worker who raises a difficult concern.

Context can help us understand behaviour, but it doesn't excuse harm. Fatigue, pressure and old patterns may explain why someone reacted badly. The behaviour still needs to be addressed, and the person still needs to answer for its impact.

3. The capability of the whole workforce

The third piece is where the evidence is least developed and where the opportunity may be greatest.

Workplaces often assume culture flows downward from leaders. In reality, culture is also created sideways.

It's created between co-workers, between experienced employees and new starters, between the person who makes the joke and the people who laugh, between the worker who raises a concern and the colleague who rolls their eyes, and between the person struggling and the co-worker who notices but does nothing.

Peers appear throughout the research on psychological safety, voice and support, yet they often disappear from the training budget.

A recent review of resilience interventions in public-sector workplaces found that multi-session programs sometimes improved resilience outcomes, with 23 of 51 calculated effects reaching statistical significance.⁸ But the evidence was highly varied, the risk of bias was often moderate or serious, and almost every intervention focused on the individual rather than the conditions surrounding them.

That's an important limitation.

We don't yet have strong evidence that teaching workers self-awareness, emotional regulation or communication skills will reliably reduce incidents or transform organisational culture. The research hasn't tested these combined capability models well enough, particularly in industries such as construction, mining, resources and contracting.⁹

That means providers need to be honest.

Workforce capability shouldn't be sold as a proven cure for psychosocial risk. It should be developed carefully, piloted responsibly and evaluated against meaningful outcomes.

But thin evidence isn't the same as evidence of failure. It may also show us where the research has been looking.

Many resilience programs have measured whether an individual's resilience score improved. Far fewer have examined whether a worker becomes more capable of recognising their internal state, interrupting an automatic reaction, communicating through conflict or speaking when something doesn't feel right.

Those are different outcomes, and they're deeply relevant to what happens in real workplaces.

What happens before someone stays silent

A worker sees something that doesn't look right.

They might feel tension in their body as their mind calculates whether they'll be dismissed, embarrassed, blamed, labelled difficult or told to stop holding up the job. They might technically know the reporting process, but knowledge isn't the only force shaping their next action.

A multitude of factors may influence what they do next. Their previous experiences may be shaping what they expect will happen. Fatigue may be narrowing their ability to think flexibly. The urgency of the task may be competing with the uncertainty they feel, or their position in the hierarchy may lead them to assume that silence is the smarter option.

This can happen in seconds.

There may be no deliberate internal conversation and no careful weighing of policy against procedure, only a rapid human response shaped by state, history, relationship and context.

A 2025 meta-analysis involving nearly 35,000 workers found that employee silence was moderately associated with burnout, and more strongly than employee voice was associated with lower burnout.¹⁰ Most of the studies were cross-sectional, so we can't confidently say whether silence drives burnout, burnout drives silence or both reinforce one another.

But the relationship matters.

Silence isn't an abstract culture score, it's a behaviour. It happens when a human being notices something and decides, consciously or unconsciously, not to say it.

Organisations can create safer systems and more supportive leaders, but people still need the internal and interpersonal capacity to act within those conditions.

An individual needs to recognise when their physical or emotional state may be affecting their perception. They need enough regulation to remain present under pressure, communication skills that work beyond the training room and the confidence to ask questions without assuming that uncertainty is weakness. They also need the humility to receive feedback, the critical thinking to challenge familiar assumptions and the self-awareness to interrupt patterns that might otherwise run automatically.

This is why The Human Impact Project works to understand the human being beneath the behaviour, as well as the internal state shaping what happens next.

Where individual capability becomes harmful

There's a version of workforce capability training that can deepen psychosocial risk rather than reduce it.

If an organisation teaches workers resilience under pressure while leaving an unsafe roster, an unmanageable workload or a supervisor who punishes honesty exactly as they are, it hasn't controlled the hazard. It has transferred responsibility for that hazard onto the people with the least authority to change it.

It may even leave the worker believing their distress reflects a personal failure to cope rather than a reasonable response to harmful conditions.

In that situation, capability training functions less as prevention and more as documentation. The organisation gets a training record, while the worker gets another expectation to manage.

Research on psychosocial safety climate shows that management priorities sit upstream of many workplace demands. In a longitudinal study of education workers, organisational climate predicted later work pressure, emotional demands, distress and exhaustion.¹¹

What senior leaders genuinely prioritise influences the conditions people experience.

Tools such as the Psychosocial Safety Climate scale can help organisations assess whether workers believe psychological health is genuinely valued, communicated and protected.¹² But measurement is only a diagnosis. A better score doesn't automatically mean the underlying work has improved.

Individual capability can't compensate for structurally harmful conditions, but structurally sound conditions don't automatically create human capability either.

A well-designed job can still be occupied by people who avoid conflict, misread one another, shut down under pressure or repeat patterns they've never been taught to recognise.

The solution isn't to choose between system change and individual development. It's to recognise that one can't substitute for the other.

What the evidence allows us to say

Changing harmful work conditions matters.⁴⁵ So does leadership behaviour.¹²³

Psychological safety supports learning, information sharing, engagement and performance, although much of the evidence remains correlational.²⁷ Individual interventions can also support wellbeing for some people, particularly when they sit alongside structural change rather than replacing it.¹³

The evidence is much thinner when it comes to workforce-wide capability programs and their effect on incidents, culture and long-term psychosocial outcomes.

A systematic review in the construction industry found only four organisational intervention studies involving 260 workers, with no robust effects demonstrated on general mental health or emotional exhaustion.⁹ That doesn't prove organisational interventions don't work. It shows how little high-quality research has been conducted in the environments that may need it most.

The same caution applies to leadership style research. Transformational leadership is generally associated with better wellbeing, but much of the evidence is observational and relies on self-reported measures.¹⁴

The responsible position isn't to promise culture transformation through a workshop. It's to acknowledge what we know, identify what we don't and build better evidence through well-designed pilots.

That also means measuring more than whether people enjoyed the training. The meaningful metrics emerge through post-training observation and follow-up:

• Did employees become more willing to raise concerns?

• Did supervisors respond differently when concerns were raised?

• Did conflict escalate less often?

• Did teams report greater trust?

• Did near-miss reporting change?

• Did people seek help earlier?

• Did the demands and conditions of the work improve?

• Did those changes last?

If we want a stronger evidence base, organisations and providers need to measure honestly and be willing to publish what didn't work as well as what did.

What this changes on Monday

The practical question isn't whether organisations should continue developing leaders. They absolutely should.

The better question is what they're entitled to expect from that development and what needs to happen alongside it.

What conditions are shaping behaviour?

Look beyond individual coping and examine workload, fatigue, staffing, role clarity, autonomy, rostering, procedural friction and the consequences people experience when they raise concerns.

The organisation should be asking whether the design of the work supports safe behaviour or quietly makes that behaviour harder.

What behaviours are leaders reinforcing?

Leadership development shouldn't be measured only by attendance or knowledge. The more revealing measure is what happens when pressure rises.

Do leaders become curious or defensive? Do they invite challenge and then punish it? Do they recognise effort while continuing to reward chronic overwork? Do they address harmful behaviour consistently, regardless of a person's seniority or technical performance?

The answers to those questions reveal what the culture is actually teaching.

What capability does the wider workforce have?

Employees also need to be equipped to navigate the human moments that policies can't manage for them.

Can they recognise when their state is affecting their perception, communicate clearly under pressure and receive feedback without immediately becoming defensive? Can they challenge a decision respectfully, notice when a colleague is struggling and pause before repeating a pattern that causes harm?

This doesn't shift responsibility away from the organisation. It recognises that culture is lived through people.

The space between the boardroom and the shift

Three hundred people are working while fourteen were in the leadership room.

Whatever happens for the shift worker at four in the morning won't be determined by a policy alone. It will be shaped by the conditions of the work, the behaviour leaders have normalised and the internal and interpersonal capabilities of the people in that moment.

Culture isn't created in the training room. It's created, reinforced or interrupted through thousands of human interactions across every shift.

The state of the individual shapes the quality of the interaction, and enough interactions shape the culture of an organisation.