Building Psychological Safety: Edmondson's Research Made Practical

The Most Replicated Finding in Team Science

In 2015, Google published the results of Project Aristotle, a two-year study examining what makes teams effective. After analyzing 180 teams across the company, tracking multiple performance metrics, and testing dozens of variables, they arrived at a clear conclusion: psychological safety was by far the most important factor (Duhigg, 2016).

Teams where members felt safe to take interpersonal risks - asking questions, admitting mistakes, proposing unconventional ideas, challenging the status quo - consistently outperformed teams where members played it safe.

Google's findings were striking, but they weren't new. They confirmed what Harvard Business School professor Amy Edmondson had been researching since the mid-1990s. Edmondson's original work began with a surprising observation: hospital nursing teams with better leadership reported more medication errors, not fewer (Edmondson, 1999). The explanation was counterintuitive: these teams weren't making more mistakes - they were more willing to report them. Their leaders had created conditions where admitting errors was safe, which made learning and improvement possible.

Edmondson defines psychological safety as "a shared belief held by members of a team that the team is safe for interpersonal risk-taking" (Edmondson, 1999). It's not about being comfortable or unchallenged. It's about being able to be candid without fear of punishment, humiliation, or marginalization.

What Psychological Safety Is Not

The concept has been widely adopted and frequently misunderstood. A few clarifications:

It's not niceness. Psychologically safe teams are often blunt. They disagree openly, challenge each other's thinking, and point out problems directly. The safety lies in the consequences of candor, not in the absence of it. In fact, teams that are "nice" in the sense of avoiding all friction often have low psychological safety - people are being polite because they don't feel safe being honest.

It's not the absence of accountability. Edmondson (2019) explicitly addresses this misconception: psychological safety and high performance standards are independent dimensions. The optimal zone - what she calls the "learning zone" - combines high psychological safety with high accountability. Low safety plus high standards produces anxiety. High safety plus low standards produces comfort but not performance.

It's not an individual trait. Psychological safety is a group-level condition. The same person might feel psychologically safe in one team and unsafe in another, depending on the norms, the leader's behavior, and the team's history. This means it can't be hired for - it has to be built.

It's not permanent. Psychological safety can be eroded by a single incident - a leader punishing someone for raising a concern, a team member being ridiculed for an idea, a reorganization that signals that candor is dangerous. It requires ongoing maintenance, not a one-time intervention.

Why It Matters: The Mechanism

The connection between psychological safety and performance operates through several mechanisms:

Information flow: In unsafe teams, people withhold bad news, concerns, and dissenting opinions. Leaders make decisions based on filtered information, which degrades decision quality. In safe teams, information flows freely, and leaders operate with a more accurate picture of reality.

Learning speed: Teams that can openly discuss failures, near-misses, and mistakes learn faster than teams that bury or ignore them. Edmondson's research in healthcare settings showed that psychologically safe teams had significantly higher rates of learning behaviors - experimentation, reflection, and knowledge-sharing (Edmondson, 2003).

Innovation: Generating novel ideas requires proposing things that might not work. In unsafe teams, people self-censor, proposing only ideas they're confident will succeed. In safe teams, people propose riskier, more creative ideas - some of which fail, and some of which produce breakthroughs.

Engagement: Gallup's research consistently shows that people who feel their opinions count at work are significantly more engaged than those who don't. Psychological safety is the mechanism through which "opinions count" becomes operationally true.

Retention: In exit interviews and engagement surveys, the most common reasons for voluntary departure aren't compensation - they're feeling undervalued, unheard, or unable to be authentic at work. These are all symptoms of low psychological safety.

Measuring It

Edmondson's original seven-item survey remains the gold standard for measuring team psychological safety. The items include:

• "If you make a mistake on this team, it is often held against you" (reverse-scored)
• "Members of this team are able to bring up problems and tough issues"
• "People on this team sometimes reject others for being different" (reverse-scored)
• "It is safe to take a risk on this team"
• "It is difficult to ask other members of this team for help" (reverse-scored)
• "No one on this team would deliberately act in a way that undermines my efforts"
• "Working with members of this team, my unique skills and talents are valued and utilized"

Scoring is simple (7-point Likert scale averaged across items and team members), and the measure has been validated across dozens of studies in multiple countries and industries.

The value of measuring isn't the number itself - it's the conversation the results enable. When a team discusses its psychological safety scores openly, that discussion is itself a psychological safety intervention.

Building It: Leader Behaviors

Edmondson's research and subsequent studies identify specific leader behaviors that predict team psychological safety:

Frame Work as a Learning Problem

When leaders frame the work as requiring input from everyone - "This problem is complex enough that no one person has the answer" - they create a rational basis for speaking up. When they frame work as execution - "We know what to do, just do it" - they signal that questioning or divergent thinking is unwelcome.

Practical language: "We're going to encounter things we didn't anticipate. When you see something that doesn't look right, I need you to flag it - that's how we get better."

Model Fallibility

Leaders who openly acknowledge their own mistakes, uncertainties, and knowledge gaps set a norm that being wrong is acceptable. "I got that wrong. Here's what I've learned" is one of the most powerful statements a leader can make.

This doesn't mean performing weakness or self-deprecation. It means honest acknowledgment of the reality that everyone, including the leader, operates with incomplete information and makes errors.

Research by Owens and Hekman (2012) on "expressed humility" in leadership found that leaders who admitted mistakes and acknowledged limitations had teams with higher levels of learning behavior and engagement.

Ask Genuine Questions

Not rhetorical questions or leading questions - genuine ones where you don't already know the answer and are visibly interested in the response.

"What are we missing?" "What's the biggest risk to this plan that we haven't discussed?" "Who disagrees with this direction and can articulate why?"

The question signals that the answer matters. But the question alone isn't enough - how you respond to the answer determines whether the question gets taken seriously next time.

Respond Productively to Input

This is the behavior that matters most, and it's the one most leaders underestimate. The moment someone takes the risk of offering a dissenting opinion, admitting a mistake, or raising a concern, the leader's response determines whether that behavior happens again.

Productive responses (these build safety):

• "Thank you for raising that. Tell me more about what you're seeing."
• "That's an important concern. How should we address it?"
• "I hadn't considered that perspective. Let's incorporate it."

Destructive responses (these erode safety, sometimes subtly):

• "We've already discussed this." (dismissal)
• "That's not really a priority right now." (minimization)
• A sigh, eye roll, or visible impatience (non-verbal shutdown)
• No response at all (the concern disappears into silence)

A single destructive response can undo weeks of safety-building behavior. People are exquisitely attuned to punishment signals, and they calibrate their risk-taking accordingly.

Create Structural Opportunities for Voice

Don't rely solely on organic contribution. Some people won't speak up in large group settings regardless of how safe the environment is - not because they feel unsafe, but because of personality or cultural norms.

Round-robins: Going around the table ensures every voice is heard, not just the loudest. Pre-meeting input: Asking for written thoughts before a discussion gives introverts time to formulate their thinking and ensures that early speakers don't anchor the conversation. Anonymous input channels: For particularly sensitive topics, anonymous surveys or suggestion mechanisms lower the barrier to candor. Small-group breakouts: People are more likely to speak candidly in groups of 2-3 than in groups of 10+.

Building It: Team Practices

Psychological safety isn't solely the leader's responsibility. The whole team contributes to the climate:

Establish Explicit Norms

Don't assume everyone shares the same expectations about how disagreement, mistakes, and conflict should be handled. Discuss it explicitly: "As a team, how do we want to handle it when someone makes a mistake? When two people disagree? When someone has a concern about the direction we're heading?"

Making these norms explicit does two things: it creates shared expectations (which reduce uncertainty), and the act of discussing norms is itself a demonstration of psychological safety.

Practice Micro-Moments of Safety

Psychological safety is built through small, repeated interactions - not grand gestures. Responding thoughtfully to a question in a meeting. Following up on someone's concern. Crediting a colleague's idea. Saying "I don't know" without embarrassment.

These micro-moments accumulate into a team climate over weeks and months. They're also the most vulnerable to erosion - one team member who consistently dismisses others' input can undermine the safety that the rest of the team has been building.

Address Violations Directly

When someone's behavior undermines psychological safety - interrupting, dismissing, mocking, or retaliating - it needs to be addressed. If the leader doesn't address it, the silence is interpreted as endorsement.

The address doesn't have to be public or punitive. A private conversation is often more effective: "I noticed that when Priya raised her concern in the meeting, your response felt dismissive. I don't think that was your intent, but I want to flag how it might have landed."

The Long Game

Building psychological safety is a sustained effort, not a project with a completion date. Edmondson's research shows that the highest-performing teams maintain their psychological safety over time through continuous attention to the behaviors and norms that create it.

The investment pays off substantially. Teams with high psychological safety not only perform better on task metrics - they also report higher job satisfaction, lower burnout, stronger commitment, and lower turnover. The mechanism is straightforward: people who can bring their whole selves to work, without performing or self-censoring, have more energy available for the work itself.

If you're a leader or team member looking to build your own capacity for creating safety in groups, developing the underlying EQ skills - self-regulation, empathy, communication, conflict management - provides the behavioral foundation that safety requires.

Psychological safety isn't a warm feeling. It's a performance strategy grounded in decades of evidence. And it starts with the question every leader and team member can ask themselves: "What happens on this team when someone takes a risk?"