When Amy Edmondson set out to study hospital teams in the 1990s, she expected to find that the best teams made the fewest mistakes. She found the opposite. The better teams, as measured by a team diagnostic survey, reported higher error rates, not lower ones.
That result didn't make sense until she looked closer. The strongest teams weren't making more mistakes. They were more willing to talk about them. As she put it, maybe better teams aren't more error-prone. Maybe they're more able to discuss their errors openly.
She named the variable that explained the difference: psychological safety. It's the confidence that you can speak up, ask a question, admit a mistake, or disagree without being punished, embarrassed, or sidelined. And it is the most reliable predictor of team performance we have.
Let me be clear about what this is not. Psychological safety is not about being nice. It is not about making work comfortable or avoiding hard conversations. That misunderstanding is where most leaders go wrong. It is about whether it is safe to tell the truth, especially to someone more powerful than you. A team can be warm and polite and still be terrified to say what actually needs to be said.
The stakes are not abstract. When Google ran Project Aristotle, a multi-year study of over 180 of its own teams, the researchers expected team composition to explain why some teams outperformed others. Instead they found that who was on the team mattered far less than how the team worked together, and that psychological safety was the single most important of the five dynamics that made teams effective. Above talent. Above seniority. Above experience.
The good news is that psychological safety is buildable. This article will act as a starting guide for leaders.
The Tale of Two Risks
Every person on your team runs a silent calculation before they speak. They weigh two risks against each other.
The first is performance risk: "If I stay quiet and we get this wrong, we underperform."
The second is interpersonal risk: "If I speak up, I might get in trouble, look foolish, or make an enemy."
Most of the time, in most organizations, the interpersonal risk wins. The cost of speaking up feels immediate and personal. The cost of staying quiet is diffuse and shared. So people swallow the idea, sit on the concern, and let the moment pass.
Psychological safety is what happens when you lower the interpersonal risk enough that speaking up becomes the obvious choice. That is the entire job. Reduce the interpersonal risk, and your team hands you every idea, every concern, and every early warning it has. That is the whole game.
Edmondson saw this play out in operating rooms. On the best teams, a nurse who spotted a better way to do something said so, and the team adopted it. On weaker teams, people saw the same opportunities and stayed silent because it felt safer not to stick their necks out. Same skill. Same information. Completely different outcome. The only variable was whether it felt safe to speak.
You have some version of this on your team right now. People are seeing things you are not. The only question is whether they are telling you.
The Minimum Viable Design: Three Moves
People have written excellent books on this, and Edmondson's "The Fearless Organization" is the place to start if you want to go deep. But if you only do three things, do these, and do them in this order.
Move 1: Set the stage. Frame the work clearly. What are we doing, why does it matter, and what are the stakes of getting it wrong? Attach a noble purpose to it, something bigger than the task itself. This is easy in healthcare and harder in other industries, but it is always worth the effort, because purpose is what makes the rest of the work feel worth speaking up for.
Setting the stage is not a one-time event. You do not frame the work once at kickoff and consider it handled. You repeat it weekly, at minimum. When teams fall apart, it is often because they framed the purpose once at the beginning and never returned to it.
Move 2: Invite participation. You have to actively pull people in, and there are three ways to do it well.
- Lead with situational humility: "I got this far on my own. I need your help for the rest." That signals you cannot do it alone and you genuinely want them in it.
- Use inquiry, asking open-ended questions that warm up people's thinking rather than fishing for the answer you already have. "When this happened, what do you think was going on?" works better than any question with a right answer.
- Use explicit process to orchestrate who speaks. Skip the dreaded "let's go around the room," which drains energy and turns participation into an obligation. Instead, conduct the conversation. Notice who has spoken and who has not, and deliberately open the door: "Does anyone see this differently?"
Move 3: Respond productively. This is where most well-intentioned leaders fall short. They set the stage. They invite participation. And then they fumble the response, which quietly undoes everything they built.
Responding productively has three parts, and they matter enough to take one at a time.
Respond Productively, Part One: Reward the Act of Speaking
When someone contributes, appreciate the fact that they spoke before you evaluate what they said. Thank the first person to speak. Acknowledge the contribution even when the idea is weak, because in that moment you are not judging the idea, you are rewarding the act of participating.
This is hard. When someone says something off-base, the instinct is to correct it immediately. Resist that. If you speak and I judge you harshly in front of the room, I have just closed the mouths of everyone else watching. They have learned that speaking up is dangerous, and they will stop. The time for judging the idea comes later. The time for appreciating the contribution is now. Speaking, like silence, is contagious.
Respond Productively, Part Two: The Failure Rule
How you respond to failure is the single biggest determinant of whether your environment is safe or unsafe.
A first-time, intelligent failure, a smart attempt at something new that did not work, should be treated almost like a success. Celebrate the learning, because the team just discovered something that no one now has to repeat. That is the right kind of wrong. It is the cost of being a team that experiments and improves.
A repeated, preventable failure is different, but the response is still not to blame the person. A failure that keeps happening is a process problem, not a people one. The system was never set up to root out the error, so the work is to go fix the system, not to find someone to punish.
When leaders punish first-time failures, they do not stop the failures. They drive risk-taking underground. People stop trying new things, and worse, they start hiding their mistakes. You lose the experiments that would have moved you forward, and you lose your early warning system, because the problems that get hidden are exactly the ones you will read about later, after they have grown.
Respond Productively, Part Three: Sanction Bad Behavior
If there is genuinely bad behavior, yelling, bullying, belittling, talking over people, it is the leader's job to step in and sanction it. Every time.
This is the most under-discussed part of psychological safety, and it is essential. If you let bad behavior slide because the person is talented or senior, you have not just tolerated it. You have endorsed it. You have emboldened the person who did it, and you have signaled to everyone else that this is how high-status people are allowed to act. The behavior spreads, and your best people go quiet or leave.
Sanctioning is not a soft skill. It is the price of keeping the environment safe for everyone else. A leader who invites participation and then fails to protect the people who participate has built nothing.
Reinforce It
Once the foundation is in place, two practices make it durable.
Assign the skeptic. Give someone the explicit job of pushing back and asking the hard questions. This is red-teaming, borrowed from how software teams deliberately hunt for vulnerabilities in a system. When you make truth-telling a formal role, you prove that it is safe, and you take the burden off any one person to be "the difficult one." Dissent becomes a job rather than a risk.
Run real postmortems. After any significant piece of work, run an After Action Review that examines two things (Substance & Style), not one. Substance is what you learned about the work itself. Style is how well you set the stage, invited participation, and responded. Most teams review only the substance. The teams that review both improve faster, and they tend to stay out of the headlines, because they catch the small issues before they compound into the big problems.
Repair It When It's Already Broken
Own it. Offer a specific apology that names what happened and the cost it caused. Not a vague "mistakes were made," but a clear acknowledgment of the harm.
Fix it. Name the procedures that are changing and how you will know they are working.
Show it. Demonstrate the new world in visible action. This is where real, visible sanctions for bad behavior matter most, because people need to see the change, not just hear about it.
Here is the encouraging part. People are wired for redemption. They want the place to work, and they want to do well in it. Expect roughly 80% of people to march forward with you once they see a genuine repair. This is often a one-week problem, not a one-year one. It is far more fixable than it feels in the moment.
What Psychological Safety Looks Like in Practice
Scenario 1: Setting the stage in a recurring meeting.
Ineffective: A team lead frames the project's purpose at kickoff and never mentions it again. Three months in, the team is going through the motions. The work is getting done, but no one remembers why it matters, and the energy has drained out of the room.
Effective: The same lead re-anchors the purpose at the top of every weekly meeting. It takes ninety seconds. The team stays connected to why the work matters, and engagement holds because the stakes never go fuzzy.
Scenario 2: Responding to a first-time failure.
Ineffective: A direct report tries a new approach to a problem. It fails. The leader reacts with visible frustration and a pointed "why did you do it that way?" The report learns the lesson immediately, and so does everyone watching: do not take risks. The next good idea never gets voiced.
Effective: The leader treats the intelligent failure as exactly what it is, valuable information the team did not have before. They surface what it taught everyone and thank the report for the attempt. Risk-taking stays alive, and the next idea comes faster because people know a smart swing-and-miss is safe.
Scenario 3: Sanctioning bad behavior on a cross-functional project.
Ineffective: A senior, talented contributor routinely talks over quieter teammates and belittles ideas that aren't theirs. The leader lets it go because the person delivers. Within a quarter, the quieter voices have stopped contributing entirely, and two strong people have started looking for the exit.
Effective: The leader addresses the behavior directly and visibly, making clear that talent does not license disrespect. The team sees that status does not buy a pass, and the people who had gone quiet start speaking again. Safety is restored because it was visibly defended.
The Stakes
Every silence in your organization is a person who decided the interpersonal risk outweighed the performance risk. Somewhere on your team, right now, someone is sitting on an idea, a concern, or a warning because speaking up feels too costly.
The leader's job is to flip that math. And it is more buildable than most people think. Set the stage and keep resetting it. Invite participation on purpose. Respond in a way that makes the next person want to speak.
Reduce the interpersonal risk, and your team hands you every idea it has.
Want the one-page version to keep at your desk? Complete our website contact form and we’ll send you the Psychological Safety Leader's Playbook, a single-page field guide you can put to use the same day.