Psychological Safety at Work: What Edmondson Found

People won't admit mistakes unless it's safe to fail.

This isn't just about creating a comfortable environment. It's the core finding from two decades of research on psychological safety—research that explains why some teams learn and improve while others stagnate.

Amy Edmondson's work, validated across industries and extended by hundreds of subsequent studies, reveals something counterintuitive: high-performing teams don't make fewer errors. They report more of them.

The difference is psychological safety. And most performance management systems systematically undermine it.

What Psychological Safety Actually Is

Let's start with what it isn't. Psychological safety isn't about being nice. It isn't the absence of conflict or accountability. It isn't a comfortable environment where no one is challenged.

Edmondson defines psychological safety as "a shared belief that the team is safe for interpersonal risk-taking." It's the confidence that you won't be punished or humiliated for speaking up with ideas, questions, concerns, or mistakes.

Think about your own experience. Have you ever held back a question because you didn't want to look ignorant? Avoided raising a concern because you weren't sure how it would land? Stayed quiet about a mistake, hoping no one would notice?

That's the absence of psychological safety—and it's more common than we might expect, even in strong organizations with talented people.

What Edmondson Found

Edmondson's original 1999 study examined 51 work teams in a manufacturing company. She expected to find that better teams made fewer errors. Instead, she found the opposite: teams with higher psychological safety reported more errors.

The insight: these teams weren't making more mistakes. They were more willing to surface them. And that willingness to surface problems early—before they compounded—led to better learning and better performance.

Edmondson's model proposes: Psychological safety → Team learning behavior → Team performance. Longitudinal and correlational evidence supports this sequence, though establishing clean causality in organizational research is notoriously difficult.

Subsequent research has reinforced and extended this finding. Frazier and colleagues' 2017 meta-analysis confirmed that psychological safety is positively related to work engagement, task performance, organizational commitment, and job satisfaction. Edmondson and Bransby's 2023 review of 185 papers found the framework holding up across contexts.

The mechanism is straightforward. Learning requires trying new things, which means risking failure. If failure carries social or professional consequences, people minimize risk. They stick with what's safe. They hide struggles rather than surface them. And the organization loses its capacity to learn.

How PM Systems Undermine Safety

Here's where performance management enters the picture—often destructively.

Many PM systems put development conversations "on the record." What you share about your growth areas, your struggles, your experiments that didn't work—all of it becomes part of your documented history. It can inform your rating. It can affect your compensation. It can shape how people perceive your capability.

In that environment, rational people edit themselves. They highlight wins. They minimize challenges. They perform rather than reveal.

This isn't a trust problem with managers. Many managers are trustworthy. It's a system design problem. When the structure creates consequences for vulnerability, people protect themselves. That's a natural response to the structure they're in.

The result: development conversations become performative. Employees share the growth areas that are safe to share—the ones that sound good, that show self-awareness without revealing real weakness. The genuine struggles stay hidden. And the development that matters most never happens.

The Safety Requirements for Learning

Edmondson's research points to what genuine development requires:

Interpersonal risk must be decoupled from professional consequences. People need to know that admitting a struggle won't be used against them. Not as a policy promise—as a structural reality.

Leader behavior matters. Managers who acknowledge their own mistakes, who respond constructively when others surface problems, who ask questions rather than assign blame—these behaviors build safety. Managers who punish the messenger, who respond defensively to concerns, who treat mistakes as character flaws—these behaviors destroy it.

Consistency over time. Safety isn't built in a single conversation. It accumulates through repeated experiences of vulnerability without punishment. One "gotcha" can undo months of trust-building.

Structural support. Individual managers can create pockets of safety through relationship skill. But sustainable safety requires systems that support it—not systems that undermine it.

What This Means for PM Design

If psychological safety is prerequisite for learning, and learning is prerequisite for development, then performance management systems need to create conditions for safety—not erode them.

This has specific design implications:

Genuinely private development spaces. Not "confidential" (which means someone could access it). Not "between you and your manager" (which means it's evaluative). Actually private—architecturally protected from evaluation processes.

Separation of development and evaluation conversations. When the purpose is growth, the stakes need to be removed. When the purpose is evaluation, the criteria need to be clear. Combining them compromises both.

Evidence-based evaluation. When assessment is grounded in documented evidence against pre-defined criteria, it's less dependent on what employees choose to reveal. This reduces the penalty for honesty in development conversations.

Manager development. Safety is created through behavior, not policy. Managers need to understand how their responses shape safety—and develop the skills to build it.

Try This

In your next one-on-one, try an explicit framing: "This conversation is about your development. Nothing we discuss here will affect your rating or be documented in your review. I want to understand where you're genuinely struggling so I can help."

Then notice: does the conversation change? Do you hear about challenges that never surfaced before? Does the person seem more or less guarded than usual?

The difference reveals the cost of combining development and evaluation—and the opportunity of separating them.