What is Psychological Safety?

Psychological Safety In The Workplace

"The belief that one will not be punished or humiliated for speaking up with ideas, questions, concerns, or mistakes, and that the team is safe for interpersonal risk taking" Amy Edmondson, 1999

Psychological safety is an emergent property of a group: not a personality trait, a mood, or a management technique. When people within a group possess psychological safety, they can confidently predict that others will react positively when they speak up, ask for help, admit a mistake, or challenge an idea. That positive predictability is what makes the risk feel worth taking. The better we can predict how a group will respond to us, the more willing we are to be honest within it. The opposite of that predictability — ambiguity about norms, expectations, or purpose — is one of the most reliable ways to silence people.

Edmondson’s definition has become the most widely cited, but it builds on earlier work. Carl Rogers introduced the term in the 1950s, Schein and Bennis built on it in the 1960s, and William Kahn first used the term in an organisational context in 1990, describing the conditions under which people could bring their full selves to work. Before either of them, the concept had been developing independently across multiple applied fields: aviation safety researchers studying cockpit dynamics in the 1970s and 80s, W. Edwards Deming writing about driving out fear from organisations, Toyota developing the Andon Cord so any worker could stop the production line without fear of consequence. None of them used the phrase “psychological safety”, but they were all solving the same problem. This convergent evolution matters, because it means there is far more evidence about what works than a casual reading of the field suggests. Psychological safety is not a new idea. Read the full history here.

One final clarification before we go further: psychological safety is not the same as trust. Trust is a belief one person holds about another. Psychological safety is a belief about the group — about what it means to be a member of this particular team, in this particular context. You can trust your colleagues individually and still not feel safe to speak up in the group. The two are related, but they are not the same thing.

What Psychological Safety Is Not

This is worth addressing directly, because the concept is frequently misrepresented — usually accidentally, but sometimes in the service of dismissing it.

Psychological safety is not about making people comfortable. A psychologically safe team is not a team without conflict or challenge; it is a team where conflict and challenge can happen honestly, without fear of humiliation or retaliation. High standards and psychological safety are not in tension. In fact, the evidence consistently shows they reinforce each other: people are more likely to hold themselves and others to high standards when they feel safe enough to name problems rather than hide them.

Psychological safety is not niceness, and it is not the absence of accountability. It is not a guarantee that there will be no consequences for poor work or poor behaviour. It is the guarantee that you will not be punished or humiliated for being honest — for raising a concern, admitting you don’t know something, or questioning a decision.

It is also not something that leaders simply bestow on teams from above. The leader-centric framing — in which a manager creates psychological safety for their people — is a subtle version of the problem it claims to solve. Psychological safety operates vertically, between people of different status, and horizontally, between peers. Both matter. And the responsibility for building it is shared.

Finally: you cannot have too much psychological safety. Where research has suggested that high psychological safety produces unwanted team dynamics, the problem is almost always a lack of shared standards and expectations — not an excess of safety. We address that argument in more detail here.

Why People Don’t Speak Up: The Calculus of Voice

Understanding what psychological safety is requires understanding why, in its absence, people stay silent even when they have something important to say.

Speaking up is rarely a simple decision. It involves a rapid, often unconscious calculation: is what I’ve noticed actually worth raising? And if it is, what will happen to me if I raise it? These are two distinct barriers, and conflating them leads to the wrong interventions.

The first is an ambiguity failure. In many organisations, people aren’t sure whether what they’ve noticed is genuinely significant, whether it falls within their remit to raise, or whether others have already seen it. Ambiguity about norms, roles, and expectations makes it rational to stay quiet — not because people fear punishment, but because they genuinely cannot read the situation well enough to know whether speaking up is appropriate. This is why clarity about expectations and group norms is one of the most practical levers for building psychological safety.

The second is a valence failure. People can see clearly that something is wrong and still choose not to speak up, because the personal cost of doing so feels too high. This is where power gradients become critical. The steeper the hierarchy — the greater the status difference between the person with something to say and the person they need to say it to — the higher the perceived cost of speaking. Decades of work in aviation safety, where this dynamic was called “authority gradient” or “cockpit gradient”, show that power differentials reliably suppress voice, and that the consequences of that suppression can be catastrophic.

These two failures — ambiguity and valence — interact. High ambiguity raises the bar at which people attempt the valence calculation at all. In organisations with unclear expectations and steep hierarchies, both gates are closed at once. We explore this in more depth in our Calculus of Voice article.

The Benefits of Psychological Safety

When people feel safe to speak up, ask for help, admit mistakes, and challenge ideas, organisations learn faster, make fewer errors, and recover from them more quickly. New ideas surface rather than staying trapped in the heads of people who assume no one wants to hear them. Risks that would otherwise stay hidden until they become crises are named early. People stay — because working somewhere that treats honesty as a strength rather than a liability is, quite simply, a better place to work.

These outcomes are well evidenced. Google’s Project Aristotle — a large-scale study of team performance — found psychological safety to be the single most important factor distinguishing high-performing teams. The State of DevOps reports have consistently identified it as foundational to software delivery performance and organisational resilience. Our own research, including work on the relationship between job security and psychological safety, on why people foster psychological safety, and on the barriers that prevent it from taking hold, points in the same direction.

But the case for psychological safety doesn’t rest only on organisational outcomes. A team in which people cannot be honest — cannot ask for help, cannot name what they’ve seen, cannot be themselves — is asking people to spend a significant part of their lives performing safety rather than experiencing it. That is worth addressing on its own terms, independent of what it does for productivity.

Psychological Safety is Not the Same for Everyone

A team is only as psychologically safe as the least safe person on it. And the people with least safety are rarely distributed randomly.

We all bring our histories with us. Previous workplaces where speaking up had consequences. Educational and social environments that taught certain people — on the basis of gender, race, class, neurodiversity, sexuality, language, and more — that their contributions would be received differently. These are not personality quirks to be coached away; they are rational responses to real patterns of experience. Even in a genuinely safer environment, it takes time to unlearn lessons that were learned the hard way.

The mechanism here is the power gradient. Structural inequalities create steeper gradients for some people than for others, which means the personal cost of speaking up is higher for people who are already marginalised or disadvantaged. Psychological safety programmes that ignore this — that treat the team as a level playing field, or that measure average safety without attending to who is least safe — do not actually address the problem. They address the experience of the already comfortable.

This is also why psychological safety and inclusion are not separate workstreams. They are the same project approached from different angles. Research confirms what practice suggests: diversity improves team performance, but only in psychologically safe environments. Without the safety to speak, the perspectives that diversity brings cannot actually enter the conversation. (Caruso and Woolley, 2008, in Edmondson and Lei, 2014.)

Psychological Safety and Accountability

One of the most persistent objections to psychological safety is that it is incompatible with accountability — that creating an environment where people feel safe to make mistakes means never holding anyone responsible for anything.

This gets the relationship backwards.

Accountability requires honesty. You cannot hold people accountable for things you don’t know about, and in low-psychological-safety environments, the things that most need addressing are precisely the things most likely to stay hidden. The choice is not between psychological safety and accountability; it is between a culture in which problems surface early and can be addressed, and one in which they remain hidden until they are too big to ignore.

Not all mistakes are the same. Errors researchers distinguish between slips and lapses (competent people making the kinds of small errors all humans make), mistakes (decisions that seemed reasonable at the time but turned out to be wrong), and intentional violations. A culture of genuine accountability responds proportionately to these different types — supporting learning from the first two, and applying appropriate consequences to the third. A culture of blame treats all three the same, which means people learn quickly to hide the first two rather than report them.

The case of nurse Hadiza Bawa-Garba, convicted of manslaughter following the death of a patient in a severely under-resourced NHS setting, illustrates what happens when systemic failures are converted into individual culpability. The system that made the error possible remained unchanged. The individual who made an honest mistake in impossible circumstances was prosecuted. Psychological safety and just accountability are not opposites — but accountability without context is, functionally, blame. We explore this in detail in our article on accountability.

Psychological Safety and the Four Stages Model

Timothy R. Clark’s Four Stages of Psychological Safety describes a progression from Inclusion Safety (feeling safe to belong) through Learner Safety (safe to ask questions) and Contributor Safety (safe to contribute ideas) to Challenger Safety (safe to challenge others and question the status quo).

It is a useful model. The progression from belonging through learning to contributing to challenging maps onto something real about how people calibrate their willingness to take increasing interpersonal risk in a group. As a framework for thinking about where a team is, and what might need to shift, it has practical value.

It is also worth being clear about what it is not. People do not move through these stages linearly, and the stages do not exist as discrete realities in any given team. Psychological safety is contextual — someone might feel safe to challenge ideas in one meeting and not another, with one colleague and not another, on one topic and not another. Models that imply a single team-level score can flatten the variation that matters most. All models are wrong, and some are useful: this one is useful, as long as it is held lightly.

Westrum’s Cultural Typologies

Ron Westrum’s 2004 typology of organisational cultures describes how information flows through an organisation, and how that flow predicts behaviour when things go wrong. His three types — Pathological (power-oriented, messengers shot, failure leads to scapegoating), Bureaucratic (rule-oriented, messengers neglected, failure leads to justice), and Generative (performance-oriented, messengers trained, failure leads to inquiry) — map directly onto what we would now recognise as varying degrees of psychological safety.

Westrum was focused on physical safety in healthcare and aviation. The insight transfers. Organisations that want to know how psychologically safe they are can ask a simpler version of his question: when something goes wrong here, what happens to the person who names it first?

Pathological	Bureaucratic	Generative
Power oriented	Rule oriented	Performance oriented
Low cooperation	Modest cooperation	High cooperation
Messengers “shot”	Messengers neglected	Messengers trained
Responsibilities shirked	Narrow responsibilities	Risks are shared
Bridging discouraged	Bridging tolerated	Bridging encouraged
Failure leads to scapegoating	Failure leads to justice	Failure leads to inquiry
Novelty crushed	Novelty leads to problems	Novelty implemented

The Westrum organisational typology model: How organizations process information ( Ron Westrum, “A typology of organisation culture),” BMJ Quality & Safety 13, no. 2 (2004), doi:10.1136/qshc.2003.009522.)

Closing the Gap: Work as Imagined and Work as Done

Organisations run on assumptions about how work actually happens. Those assumptions are often wrong. The gap between work as imagined — how leaders and systems believe work is done — and work as done — how it actually happens at the sharp end — is one of the most reliable sources of risk, inefficiency, and avoidable harm.

Closing that gap requires honesty from the people closest to the work. And honesty from people closest to the work requires psychological safety — specifically, the kind of safety that allows someone to say “this is not how we actually do it” to someone with the authority to respond badly to that information. Without it, the gap stays open, and organisations continue to design interventions for a version of work that doesn’t exist.

Measuring Psychological Safety

Psychological safety can be measured — through surveys that ask team members to rate agreement with statements reflecting group norms, or through more discursive approaches like the psychological safety matrix, which encourages teams to describe where they are rather than simply score it. Both have their uses: quantitative measures provide a baseline and a way to track change over time; qualitative approaches often surface the texture that scores miss.

There are important considerations before measuring, however. Anonymous surveys in low-trust environments can themselves become a source of anxiety. Aggregated scores can mask the experience of the least safe people. Measurement without a genuine commitment to act on what is found can do more harm than not measuring at all. Our measurement resources go into this in more detail.

Our Research

At Psych Safety we carry out our own research on psychological safety alongside our practice work. Our studies have explored the relationship between job security and psychological safety — finding that precarious employment significantly undermines the conditions for safety — examined why people choose to foster psychological safety, and investigated the structural and organisational barriers that prevent psychological safety from taking root even where it is genuinely valued. That last body of work is particularly important, because it moves the conversation from individual behaviour change toward the conditions that make behaviour change possible.