Teams that feel safe to speak up learn more

In one line: teams where people feel safe to ask, admit mistakes and disagree tend to learn more together, and that learning goes with doing the job better.

What she did

Amy Edmondson studied 51 work teams in one large manufacturing company. The teams did different kinds of work. Team members filled in surveys about how safe it felt to speak up in their team. They also rated how often the team did “learning” things: asking for help, seeking feedback, sharing information, talking about mistakes and trying new ways of working. Managers rated how well each team performed. Then she tested whether safety, learning and performance were linked in a chain.

So this is a business study, not a healthcare one. But it grew out of her earlier hospital study, described below, and that’s the part most people quote.

What she found

• Feeling safe went with more learning. Teams that felt safer reported doing more of the learning behaviours. This held after allowing for other differences between teams.
• Learning went with better performance. Teams that did more learning were rated as performing better. Safety seemed to help performance through learning, not on its own.
• Leaders mattered. Teams felt safer where the leader coached them and where the wider company gave them support. How leaders react when people ask questions or raise problems shapes whether people keep doing it.
• The hospital finding came first (1996). In eight nursing units across two hospitals, the better-led units had higher recorded medication-error rates, not lower. An observer who didn’t know the error figures found that staff on those units were much more open about discussing mistakes. So the difference was probably in reporting, not in errors made.

How much should you trust it?

A well-known, influential link from surveys; it doesn’t prove cause, and the main study wasn’t in healthcare.

• The 1999 study measured everything at one time. Safety might lead to learning, or teams that learn well might come to feel safer.
• Most of the data are people rating their own team. People may rate their team more kindly than it deserves.
• The teams worked for a manufacturing company, not a hospital. Hierarchies, stakes and team set-ups in healthcare are different.
• The hospital study was small and exploratory: eight units. It’s a striking result, but it’s one study.

If you need to convince someone

In our words (a paraphrase of Edmondson, 1996, not a quotation): the better-led nursing units reported more errors than the others. Most likely that wasn’t because they made more errors, but because staff there felt safe enough to admit them.

What it means for you

1. Read silence carefully. If nobody asks questions, raises concerns or reports near misses, that isn’t proof all is well. It may mean people don’t feel safe to speak. Look at both: why errors happen, and whether they’re being reported.
2. Your next reaction counts. How you respond to a wrong answer, or to “I’m not sure”, shows the group what happens to people who speak up. A sigh is a response. So is silence.
3. If you want a change in practice, make it safe to practise. People need to try the new thing, get it wrong in front of you, and ask. Plan the session so that’s normal.

How it appears in Teaching That Lands

It isn’t taught directly in the three core sessions. Psychological safety, debriefing and feedback are kept for a separate session based on simulation. You’ll see small signs of it in Session 1, though. Saying whether you can read the colour-vision plates is voluntary. And feedback on each other’s slides follows a kind, two-part frame: one thing that works, one question. It’s useful background for any teaching where people need to admit what they don’t know.

The small print

• What the term means. Team psychological safety is a belief, shared by a team, that it’s safe to take risks with each other, like asking, admitting a mistake or disagreeing. The term is older than this paper: Schein and Bennis (1965) and Kahn (1990) used it first. Edmondson’s step was to make it a team idea and find a way to measure it.
• “Through” learning. Researchers call this pattern mediation: learning behaviour sat in the chain between safety and performance.
• One point in time. The design was cross-sectional: all the measures were taken at once, so cause and effect can’t be firmly shown.
• A team measured through individuals. Safety is meant to belong to the team, but it’s measured by asking individuals and combining their answers. That raises questions about how well the combined score reflects the team.
• Later work. Edmondson’s 2018 book, The Fearless Organization, pulls her work together. For studies in healthcare since, see the theory page.