Psychologically Safe
The team habit that turns mistakes into progress
A researcher named Amy Edmondson once set out to study medical errors in hospitals. Her hypothesis was straightforward: better teams make fewer mistakes. She surveyed nurses and doctors about teamwork quality, then cross-referenced those ratings with error reports.
The results were backwards.
The teams with the best teamwork scores — the ones nurses described as most cohesive, most supportive, most communicative — were reporting more errors, not fewer. The lower-rated teams, the ones with more tension and friction, appeared on paper to be making fewer mistakes.
Edmondson could have concluded she was wrong. Instead, she kept looking. She sent a covert observer into the units. And when she compared what that observer saw with what the teams reported, the picture flipped completely: “psychologically safe teams reported more errors, but they actually made fewer errors.”
The high-performing teams weren’t making more mistakes. They were admitting to more of them.
A Silence that Kills
In a psychologically unsafe environment, the incentives around mistakes run exactly backward. When people fear blame, punishment, or humiliation for surfacing problems, they hide those problems. A nurse who notices a dosage error might say nothing if she fears embarrassing a senior physician. An engineer who suspects a structural flaw might stay quiet if speaking up means questioning a decision the team has already committed to. A new employee who identifies a gap in the process might assume it was intentional, rather than risk looking uninformed.
This silence feels rational in the moment. But it’s catastrophic in aggregate.
What Edmondson found in the poorly-rated hospital teams was exactly this: errors were happening, but they weren’t being reported. Nobody was learning what caused them. Nobody could prevent them from happening again. The same mistakes recurred, quietly, invisibly, at a cost that never appeared on any report because it never had to.
The high-performing teams, by contrast, reported their errors freely. “By freely admitting their mistakes, they were then able to learn what had caused them and eliminate them, moving forward. In psychologically unsafe teams, people made fewer errors but kept repeating the same mistakes.” In psychologically safe teams, errors became information. In psychologically unsafe teams, they became secrets.
— -
What is Psychological Safety?
The term “psychological safety” was coined by Edmondson, and it’s worth being precise about what it means — because it’s frequently misunderstood.
Psychological safety is not the same as comfort. It’s not about making everyone feel good, or avoiding conflict, or softening feedback until it’s meaningless. It’s not about coddling or consensus.
Psychological safety is the shared belief that the team is safe for interpersonal risk-taking. It means people can raise concerns, admit errors, ask questions, and disagree with the prevailing view without fearing that doing so will be held against them. It’s the condition under which honest information flows.
The distinction matters because leaders sometimes confuse the two. A team where everyone agrees and no one challenges decisions isn’t psychologically safe — it’s just conformist. Psychological safety isn’t the absence of tension. It’s the presence of trust — the confidence that honesty won’t be punished.
Get Benjamin Hoyle’s stories in your inbox
Join Medium for free to get updates from this writer.
Research across industries consistently shows that psychologically safe teams outperform their peers on almost every measurable dimension: innovation, quality, efficiency, retention. A landmark Google study — Project Aristotle — examined hundreds of internal teams and found that psychological safety was the single strongest predictor of team effectiveness. Not individual talent. Not team composition. Not seniority. The degree to which people felt safe being honest.
The Leaders Role
Psychological safety doesn’t emerge spontaneously. It’s built — or undermined — by leaders, in small moments, over time.
When a manager responds to bad news by demanding to know who made the mistake, they signal that honesty is dangerous. When a senior person dismisses a junior employee’s concern without engaging with it, they signal that some voices don’t count. When a leader shoots the messenger they teach everyone in the room what happens when you speak up.
The inverse is also true. When leaders model intellectual humility — when they say “I was wrong” or “I don’t know” without apparent discomfort — they make it easier for others to do the same. When they respond to surfaced errors with curiosity rather than blame, they create a template: here is how we treat mistakes in this team.
Adam Grant connects this explicitly to the rethinking cycle. A team where people feel safe admitting errors is a team where rethinking can actually happen. If the norm is to defend decisions rather than examine them, the team gets locked into the same mental modes we explored earlier — preaching its own strategy, prosecuting dissent, politicking for internal approval. Psychological safety is the environmental condition that makes the shift to scientist mode possible.
The Relationship Conflict Distinction
There’s a nuance worth naming: not all conflict is created equal.
Research on team dynamics distinguishes between relationship conflict — friction driven by personal animosity, status disputes, and interpersonal hostility — and task conflict — disagreement about ideas, methods, and decisions. These two types of conflict have opposite effects.
Relationship conflict damages teams. It erodes trust, drains energy, and causes people to disengage. A team with high relationship conflict will underperform a conflict-free team almost every time.
Task conflict, however, can enhance performance. When teams argue genuinely about ideas — when they probe assumptions, challenge conclusions, and push back on proposals on the merits — they tend to arrive at better decisions. Grant puts it plainly: “The absence of conflict is not harmony, it’s apathy.” Relationship conflict is generally bad for performance, but task conflict can be genuinely useful — “it can be constructive. It can help groups surface their assumptions, invite curiosity about each other’s views, and promote the kind of task conflict that clears the path to rethinking.”
Psychological safety enables productive task conflict. When people trust that disagreeing won’t damage their standing, they actually disagree — and the team gets to find out whether the original idea was right.
Returning to Edmondon’s Hospitals
Edmondson’s finding, when you sit with it, is both obvious in retrospect and genuinely revolutionary in its implications. Of course teams that can talk about their mistakes learn faster. Of course teams that hide their errors keep repeating them. The logic is clean.
What makes it powerful is how thoroughly most organizations get it backward. We punish mistakes. We evaluate leaders on whether their projects succeed, not on whether their teams learn. We reward confidence and penalize doubt. We design incentive systems that make hiding failure more rational than surfacing it.
And then we wonder why we keep making the same mistakes.
The teams that perform best over time — in hospitals, in software companies, in military units, in any domain where complexity and uncertainty are real — are the ones where people can say, without fear: something went wrong. Here’s what I saw. Here’s what I think caused it. Here’s what I think we should do differently.
That conversation, more than any strategy or talent pool or technology, is what separates organizations that improve from organizations that stagnate. It’s not a cultural luxury. It’s the mechanism by which teams actually get better.
The conditions for that conversation are not complicated. They just require leaders who have decided that honest information matters more than the comfort of looking like everything is fine.
References
Grant, Adam. (2021). Think Again: The Power of Knowing What You Don’t Know. Viking Books.
