What Is Psychological Safety? The Research, Minus the Buzzword
In this article
- The real definition: a shared belief that the team is safe for interpersonal risk-taking
- What it is not: niceness, comfort, or lowered standards — the most common corporate misreadings
- The evidence: from Edmondson's hospital error study to Google's Project Aristotle
- Four leader behaviours that actually build it, and the four stages teams move through
Psychological safety may be the most successful — and most misunderstood — idea to cross from organisational research into everyday work life. It has been credited by Google with underpinning its best teams, adopted by hospitals and airlines, and printed on enough posters that many people now assume it means keeping everyone comfortable.
It doesn't. Here's the concept as the research actually describes it — and how to build the real thing.
The definition
Psychological safety, as defined by Harvard Business School professor Amy Edmondson, is:
A shared belief, held by members of a team, that the team is safe for interpersonal risk-taking.
Every word is doing work. Shared belief — it's a property of a team's climate, not an individual's confidence; the same person can feel safe in one meeting and silent in the next. Interpersonal risk — the risks in question are social: looking ignorant for asking, incompetent for admitting error, negative for raising a concern, disruptive for proposing change. In an unsafe climate, the safest personal strategy is silence — and silence is exactly what complex work cannot afford.
Where the idea comes from
The concept has a wonderful origin story. In the 1990s, Edmondson studied medication errors across hospital teams, expecting the best-led teams to make the fewest mistakes. The data showed the opposite — until she realised the best teams weren't making more errors, they were reporting more. The climate, not the competence, was the variable. Her 1999 paper formalised the idea, building on foundations laid by Edgar Schein and Warren Bennis decades earlier.
The evidence has compounded since. Most famously, Google's Project Aristotle examined more than 180 of its own teams to find what separated the best, and landed on psychological safety as the foundational factor — ahead of who was on the team at all. Meta-analyses across industries link it to learning behaviour, error reporting, innovation and performance, with the effect strongest exactly where work is complex and interdependent.
What psychological safety is not
Three misreadings account for most of the buzzword damage:
- It is not niceness. Relentlessly polite teams are often the least safe — disagreement has gone underground.
- It is not comfort. The point of safety is to make discomfort survivable: hard feedback, dumb questions, bad news early.
- It is not lowered standards. Edmondson is explicit: safety and standards are two different dials. Low safety with high standards produces anxiety and hiding. High safety with low standards produces complacency. The learning zone requires both turned up.
A useful companion model comes from Timothy R. Clark, who describes four stages teams move through: inclusion safety (you belong here), learner safety (you can ask and err), contributor safety (your work is welcome), and challenger safety (you can question how things are done). Most teams get partway; the value concentrates at the end, because challenger safety is where bad strategies get caught early.
How leaders actually build it
The research keeps returning to a handful of leader behaviours — small, repeated, unglamorous:
- Frame the work honestly. If the work is complex and novel, say so out loud. When error is expected and discussable, reporting it becomes professional rather than shameful.
- Admit your own fallibility. "I may miss things — I need to hear from you" is a sentence that redistributes permission.
- Ask genuine questions. Inquiry signals that speaking up is wanted. Teams calibrate to what leaders ask about, not what they claim to value.
- Mind the receiving moment. The response to the first piece of bad news writes the policy. Thank first, triage second. One public wince taxes truth-telling for a year.
Notice what's absent: workshops, slogans, trust falls. Psychological safety is built the way it's destroyed — in seconds, in meetings, in the moment after someone takes a risk.
Why it sits at the centre of the bigger map
On the map of fields behind this site, psychological safety anchors Belonging — but its wiring runs everywhere. It is the permission condition for curiosity: people ask the questions they can afford, and a team's intelligence is bounded by what its members are willing to say aloud. It's measurable, too — most instruments descend from Edmondson's original seven survey items, and team-level scores routinely explain why identical processes produce different results in different rooms. If you're evaluating team assessment tools, climate measures like this belong beside the personality lenses, not behind them.
And it's trainable — which is the note to end on. Psychological safety is not a personality trait of lucky teams. It's a climate, produced by observable behaviours, changeable within weeks. Of all the conditions that determine whether a group of smart people becomes a smart team, it is the one most completely in a leader's hands.
Psychological safety is one of the defining challenges FLAME addresses for workplaces — and a theme Brad speaks on for leadership teams. Explore the keynotes, or read the companion profile: Amy Edmondson Simplified.