Edmondson's Definition: A Team-Level Belief
The concept of psychological safety was defined and measured by Amy Edmondson, professor at Harvard Business School, in her 1999 paper in Administrative Science Quarterly. She defined it as a shared belief held by members of a team that the team is safe for interpersonal risk-taking. She developed the idea for a general readership in her 2018 book The Fearless Organization.
The word "shared" matters. Psychological safety is not a feeling that lives inside one person, and it is not a personality trait that some people have and others lack. It is a property of a group, something that emerges between people who work together and that can differ sharply from one team to the next inside the same organisation. Two teams reporting to the same director, sitting on the same floor, can hold completely different beliefs about whether it is safe to speak up.
The interpersonal risks Edmondson has in mind are ordinary and constant. Asking a question that might sound naive. Admitting you do not understand. Owning a mistake. Disagreeing with a senior colleague. Flagging a concern about a decision that already has momentum. In a team with psychological safety, people take these small risks routinely, because experience has taught them it is safe to do so. In a team without it, people calculate that silence is safer than candour, and they go quiet.
That silence is the real cost. The information a team needs to catch errors, improve and adapt lives in the heads of its members. Psychological safety determines whether that information is voiced or withheld.
The Four Dimensions of Psychological Safety
Psychological safety is easier to act on when it is broken into its practical components. Drawing on Edmondson's measurement work, four dimensions describe how it shows up in a team's daily behaviour.
1. Willingness to help one another
In a psychologically safe team, asking a colleague for help is normal rather than a confession of weakness. People offer support without being asked, and requesting it carries no penalty. Where this dimension is weak, people struggle alone rather than reveal a gap, and problems stay hidden until they are expensive.
2. Inclusion and diversity of perspective
Every member, including the most junior and the newest, believes their perspective is wanted. Different views are treated as useful rather than awkward. Where inclusion is weak, the same few voices dominate while others learn that contributing is not worth the social cost, and the team loses the perspectives it most needs to hear.
3. Attitude to risk and failure
Mistakes are treated as information to learn from rather than occasions for blame. People report errors and near-misses early because they expect a constructive response. Where this dimension is weak, failure is hidden, near-misses go unreported, and the team repeats the same mistakes because no one feels safe enough to name them.
4. Open conversation and the freedom to challenge
People feel able to raise difficult issues and to disagree, including with those more senior than they are. Challenge is heard rather than punished. Where this dimension is weak, decisions go unchallenged not because they are sound but because dissent feels unsafe, and obvious problems sail past unspoken.
These four dimensions are why psychological safety is so closely tied to a working speak-up culture. Each one describes a different moment where a person decides whether to voice what they know or keep it to themselves.
Why Psychological Safety Matters
The most widely cited evidence for the importance of psychological safety comes from Google. Through its re:Work programme, Google ran a multi-year internal study called Project Aristotle, published in 2015, to find out what distinguished its highest-performing teams. The study examined many plausible factors, from the seniority of members to how often colleagues socialised outside work. The qualitative finding was that, of the five dynamics that mattered, psychological safety was the most important. It sat underneath the others as the condition that allowed them to function.
The logic is straightforward once the definition is clear. A team's performance depends on the behaviours its members are willing to perform: speaking up after a near-miss, challenging a flawed plan, admitting a mistake while it is still cheap to fix. Every one of those is an interpersonal risk, and psychological safety determines whether people take it or stay silent. Without it, talent, strategy and process all underperform, because the information that would have improved the work never reaches the people who needed it.
This is why psychological safety is treated as foundational rather than as one initiative among many. In clinical teams it is linked to whether staff report safety concerns. In financial services it underpins conduct and culture. In any setting where the cost of an unspoken concern is high, the team-level belief about whether it is safe to speak determines whether the team learns or repeats. We explore the practical organisational case in detail in our guide to psychological safety training.
What Psychological Safety is NOT
Psychological safety is one of the most misunderstood ideas in organisational development, and most of the misunderstanding comes from confusing it with comfort. Edmondson herself spends considerable energy correcting this. The clearest way to understand the concept is to be precise about what it is not.
It is not niceness. A team can be unfailingly polite and have very little psychological safety, because the politeness is what stops people raising hard truths. Surface harmony often masks an absence of safety rather than the presence of it.
It is not comfort or an easy life. Psychological safety does not mean lowered expectations or protection from challenge. Edmondson pairs it explicitly with high standards. In her framing, psychological safety and accountability are two separate axes. High safety combined with high accountability is the learning zone where performance happens. High comfort with low accountability is the apathy zone, where people feel relaxed and nothing improves.
It is not the absence of conflict. Psychologically safe teams tend to have more open disagreement, not less, because people feel safe enough to voice it. The disagreement is about ideas and decisions, conducted with mutual respect, not personal attack. A team with no visible disagreement is more often a team that has gone silent than a team that genuinely agrees.
It is not a licence to underperform. Psychological safety makes it safe to admit a mistake; it does not make the mistake acceptable. The point is to surface problems early so they can be fixed, not to excuse them. Safety without standards produces complacency, which is the opposite of the outcome the concept is meant to deliver.
It is not an individual trait. No person "is" psychologically safe. Safety is created between people through repeated experience of how candour is received. That is also why it can be built deliberately, which is the subject of the next section.
How to Build Psychological Safety
Because psychological safety is a shared belief formed through experience, it is built through what leaders and teams repeatedly do, not through what they announce. Stating that "this is a safe space" tends to achieve very little. The belief updates only when people take a small risk and see it handled well. The practical levers below all work on that mechanism.
- Frame the work as a learning problem. Edmondson recommends leaders explicitly frame the work as uncertain and interdependent, so that questions, errors and concerns are understood as expected features of the work rather than as failures. When people know the work is hard and the answers are not all known, raising a problem feels normal.
- Invite participation, and ask genuine questions. Leaders set the tone by asking rather than telling, by acknowledging the limits of their own knowledge, and by actively asking quieter members what they see. Curiosity from a leader is one of the strongest signals that other voices are wanted.
- Respond productively to candour. The single most powerful builder of safety is how the first risks are received. When someone admits a mistake or raises a concern, a response of appreciation and constructive problem-solving teaches the whole team that speaking up is safe. A dismissive or punitive response teaches the opposite, instantly.
- Model fallibility from the top. Leaders who openly own their own mistakes and uncertainties make it safe for others to do the same. Safety flows downward; a leader who is never wrong creates a team that is never honest.
- Hold safety and accountability together. Building safety does not mean dropping standards. The aim is the learning zone: high expectations and high safety at the same time, so that people are stretched and also able to be honest about where they are struggling.
The reason psychological safety is genuinely hard to build is that the decisive behaviours happen in pressured, real-time moments: the meeting where a junior colleague has the right answer but hesitates, the debrief after something went wrong, the review where a difficult observation needs to surface. These are exactly the moments that conventional training, the keynote or the e-learning module, cannot reach, because they describe the behaviour instead of letting people rehearse it. This is why Sidestream's work uses immersive simulation to put leaders and teams inside those moments and let them practise the response. The principle is the same one set out in our guide to building a speak-up culture: behaviour changes through lived experience, not through being told.
How to Measure Psychological Safety
Psychological safety can be measured, which is part of why it has become a serious management concept rather than a slogan. There are two complementary approaches, and the strongest practice uses both.
The validated survey. The standard instrument is the seven-item psychological safety scale that Edmondson published with her original 1999 research. It asks team members to rate statements about whether mistakes are held against them, whether difficult issues can be raised, whether it is safe to take a risk, and similar. Because it is short, validated and widely used, it allows a team to benchmark itself and to track change over time. It is a measure of the belief.
Behavioural observation. A survey captures what people report; observation captures what they do. Behavioural measures look at the team in action: how often people speak up in meetings, whether decisions are challenged before they are made, how the team responds when someone reports an error, whether near-misses are surfaced. This is essentially measurement at Level 3 of the Kirkpatrick model, the level that assesses changed behaviour in real work rather than reported satisfaction. We explain that framework in our guide to the Kirkpatrick model.
The two are strongest together. A survey alone can be inflated by what people think they are supposed to say; observation alone lacks the team's own perspective. Sidestream's standard practice pairs the validated Edmondson scale with Kirkpatrick Level 3 observation, so the reported belief and the observed behaviour can be read against each other.
Common Misconceptions
Beyond the distinctions above, three misconceptions recur often enough to name directly.
"Psychological safety means I cannot give hard feedback." The reverse is true. Safety is what makes candid feedback possible, because it is received as help rather than attack. A team without safety is one where the hard feedback never gets said.
"We did a workshop, so we have it now." Psychological safety is not a certificate or a one-off event. It is a belief that is continually updated by how the next risk is handled. A single session can raise awareness, but the belief is rebuilt or eroded in every meeting that follows. This is why durable change depends on rehearsing the behaviour and embedding it, not on a one-day intervention.
"It is a soft, feel-good extra." Psychological safety is measurable, it is tied in the research to learning and performance, and its absence has been implicated in serious failures across healthcare, aviation and finance. It is better understood as basic operating infrastructure for any team whose work depends on people sharing what they know.
Where Sidestream Fits
This guide is the definitional reference: what psychological safety is, where it comes from, and how it is built and measured. When the question moves from understanding the concept to changing behaviour in a real team, that is the work Sidestream does.
Some things cannot be taught, they have to be felt. Psychological safety is the clearest example. You cannot lecture a team into trusting one another; the belief forms only when people experience candour being handled well. That is why our approach is built on immersive, lived rehearsal of the exact moments where safety is created or lost, measured against the Edmondson scale and observed behaviour rather than a satisfaction survey. Get in touch today. We are Sidestream.
Related Sidestream Guides
- Psychological Safety Training London, our Edmondson-anchored immersive training service
- How to Build a Speak-Up Culture, the practical companion to this guide
- What is the Kirkpatrick Model?, the framework behind Level 3 behavioural measurement
- Glossary: 100 Behaviour Change Terms
Frequently Asked Questions
What is psychological safety?
The shared belief within a team that it is safe to take interpersonal risks: to ask a question, admit a mistake, raise a concern or offer an idea without fear of being embarrassed or punished. Defined by Amy Edmondson in 1999.
Who coined the term psychological safety?
Harvard professor Amy Edmondson defined and measured the modern team-level construct in her 1999 paper in Administrative Science Quarterly, and developed it for a general readership in her 2018 book The Fearless Organization.
What are the four dimensions of psychological safety?
Willingness to help one another, inclusion and diversity of perspective, attitude to risk and failure, and open conversation including the freedom to challenge. Each describes a moment where a person chooses candour or silence.
Is psychological safety the same as being nice?
No. It is not niceness, comfort or lowered standards. Edmondson pairs it with high accountability. Psychologically safe teams have more candid disagreement, not less, because people feel safe enough to challenge each other.
How do you measure psychological safety?
With Edmondson's validated seven-item scale, ideally combined with observation of actual behaviour such as speak-up frequency and how teams respond to error. Sidestream pairs the scale with Kirkpatrick Level 3 behavioural observation.