Psychological safety at work: how to build and measure it

Psychological safety at work is the shared belief that your team is safe for interpersonal risk-taking: that you can ask a question, admit a mistake, disagree with the plan, or float a half-formed idea without fear of being punished or humiliated. Harvard’s Amy Edmondson defined the concept in its modern team sense, and it has since become one of the most cited ideas in how teams actually perform.

Most articles about it stop at the definition. This one goes further, because the gap on almost every team is not understanding the idea, it is knowing whether you actually have it. So this guide covers what psychological safety is and what it is not, why it drives performance, how to spot when it is missing, and, the part everyone skips, how to measure it on your own team and build it back up.

What psychological safety is (and what it isn’t)

The clearest way to understand psychological safety is to be precise about what it rules out, because the concept is misread constantly.

It is not about being nice. A psychologically safe team is often more direct, not less, because people trust that a blunt question or a dissenting view will be treated as useful rather than held against them. Politeness that avoids the real issue is the opposite of safety; it is a team managing its image. Edmondson’s own research makes the point that safety and high standards are not a trade-off. The teams you want are candid and demanding at the same time.

It is not comfort, and it is not lowering the bar. Safety does not mean nobody is ever challenged or that hard feedback disappears. It means the challenge lands on the work, not the person. A team can feel safe and still hold each other to a high standard, and that combination is exactly where good work comes from.

And it is a property of the team, not of an individual. It is a shared belief about how this group responds when someone takes a risk. That is why one person cannot simply decide to feel safe; the belief is built from what the team repeatedly does when a colleague speaks up, asks for help, or gets something wrong. If you want a plain-language read on how that feels day to day, our post on how to tell if your team feels psychologically safe walks through the everyday signals.

Why it makes teams perform better

The strongest evidence comes from Google. In Project Aristotle, Google set out to find what separated its effective teams from the rest, expecting the answer to be about who was on the team. It was not. The biggest differentiator was how the team worked together, and psychological safety was the single most important of the five dynamics they identified, sitting underneath dependability, structure and clarity, meaning, and impact.

The mechanism is not mysterious. On a team where speaking up is safe, problems surface while they are still small. People ask for help before they are stuck for a week. Bad news travels fast enough to act on. Someone floats the risky idea that turns out to be the good one, and someone else admits the mistake that stops it from repeating. On a team where it is not safe, all of that goes quiet, and the cost shows up later as missed risks, rework, and decisions nobody felt able to question. Safety is not a soft benefit sitting to one side of performance. It is the condition that lets a team learn fast enough to perform at all.

Signs your team doesn’t have it

You rarely get told outright that a team feels unsafe. You read it from behaviour. A few reliable signals:

• Meetings are quiet, and the same two or three voices carry every discussion.
• Nobody disagrees with the senior person in the room, and decisions pass without real challenge.
• Bad news arrives late, packaged, or only in private afterwards.
• Mistakes get hidden or explained away rather than surfaced and fixed.
• People stop asking questions, because asking has started to feel like admitting you don’t know.
• Feedback flows down the hierarchy but almost never up.

Any one of these can have an innocent explanation. Several of them together, week after week, is the pattern. The hard part is that the teams with the least safety are often the ones least able to say so out loud, which is exactly why you cannot rely on asking in the room. You have to measure it in a way that protects the honest answer.

How to measure it

Here is the gap almost every resource leaves open: plenty of people can define psychological safety, and very few give you a repeatable way to measure it. Yet it behaves like any other team signal. You can rate it, track it, and watch it move.

The instrument is a short, anonymous survey. Ask people to rate agree-or-disagree statements rather than answer open questions, so you get a comparable score and a conversation starter in the same line. A workable set, adapted from the way Edmondson’s research frames the concept:

• I can raise a problem or a tough issue on this team.
• Making a mistake here is not held against me.
• People on this team are able to bring up problems and hard questions.
• It is safe to take a risk or disagree on this team.
• No one on this team would deliberately undermine my work.
• My unique skills and contributions are valued and used.

Two rules make the difference between a real measurement and a polite fiction. Run it anonymously, because the moment people sign their name to a low score, the low scores dry up. And run it repeatedly, because a single reading is a snapshot, while a cadence gives you a trend, and the trend is what tells you whether something you changed is working.

Running the survey by hand once is easy. Running it anonymously, on a cadence, and turning the scores into a trend the team can see is where a tool earns its place. Our psychological safety health-check template does exactly this: anonymous ratings against a safety-focused question set, the spread and the trend visualised for the team, and a discussion built into the flow. It also sits inside the broader practice covered in our team health check guide, where safety is one dimension among several you track over time.

Exercises that actually build it

Safety is built by what a team does repeatedly, not by a one-off workshop, so treat these as habits to run often rather than a checklist to complete once. Each one is small, and each one signals that taking an interpersonal risk on this team is welcome.

• Run collaborative icebreakers, not performances. Choose a check-in that needs people to actually work together or contribute something of their own, not just watch one person talk. A low-stakes shared activity gets everyone participating early, which makes speaking up later feel normal rather than exceptional.
• Model vulnerability yourself first. If you facilitate, whether as a Scrum Master, team lead, or founder, go first. Name something you got wrong, or openly ask the team for feedback on how you ran the session. When the most senior person in the room shows that admitting a gap is safe, everyone else recalibrates what is allowed.
• Make asking for feedback a team exercise. Set aside a slot where the whole team both requests and gives feedback, rather than leaving it to happen ad hoc or only in one direction. Practising it as a group turns feedback from a rare, loaded event into a routine part of how the team works.
• Open with a quick check-in. A short round where everyone says one thing, even a single word on how they are arriving, gets every voice into the room in the first few minutes. It is one of the cheapest ways to stop a meeting from becoming a conversation between the two loudest people.
• Balance power, access, and contribution on purpose. Safety erodes when the same people always hold the floor. Aim for a balance of power (no single voice decides by default), a balance of access (everyone can contribute on the same channel, at the same time), and a balance of contribution (the facilitator deliberately passes the floor to people who have not spoken yet). Independent, simultaneous input beats a go-round that anchors on whoever starts.
• Offer anonymity as an option. For the topics people find hardest to say out loud, let them contribute anonymously. Anonymity is not a permanent hiding place; it is a ramp that surfaces the honest material, which the team can then discuss openly once it is on the table.
• Show a concrete win from feedback. Point to a specific example where someone raised something and it led to a real change and a good outcome. Nothing builds safety faster than proof that speaking up here actually moves things, and one vivid, true story does more than any amount of encouragement.

How retrospectives and health checks help

The exercises above share a problem: run once, they fade. Psychological safety needs a recurring container, and two rituals give it one.

A retrospective is a standing, low-stakes place to practise speaking up about the work. Run it well and it rehearses the exact behaviours safety is made of: naming what went wrong without blame, disagreeing about causes, and agreeing on a change. The facilitation choices are what make it safe. Gather input independently before any discussion so the conversation does not anchor on the first or most senior voice, keep contributions anonymous when the topic is sensitive, and make sure the floor moves around the room. We cover why the mechanics matter so much in how online retrospective tools build psychological safety.

A health check does the measuring. Where a retro works on the last sprint, a health check reads the state of the team itself, safety included, on a stable set of questions you can compare quarter to quarter. Together they close the loop: the health check tells you whether safety is trending up or down, and the retro is where you practise the habits that move it. That combination, measure it and then work on it, is the whole point of running health checks rather than hoping the culture sorts itself out.

Measure it in TeamRetro

You cannot improve what you cannot see, and psychological safety is measurable. Run an anonymous safety check, watch the trend across a few quarters, and use your retrospectives to work on whatever the scores are telling you.

TeamRetro runs the whole loop: the psychological safety health-check template for the anonymous, repeatable measurement, retrospectives to build the habits behind the numbers, and trends so you can see safety improve rather than guess at it. Start a free trial and run your first psychological safety check this week.

Keep reading

• Team health checks: how to measure how your team is really doing: the broader practice this safety check sits inside.
• Running effective retrospectives: the facilitation choices that make speaking up safe.
• How to tell if your team feels psychologically safe: the everyday signals, in plain language.

Frequently asked questions

What is psychological safety at work?

Psychological safety at work is the shared belief that a team is safe for interpersonal risk-taking: that you can ask a question, admit a mistake, disagree, or float a half-formed idea without fear of being punished or humiliated. Harvard’s Amy Edmondson defined the concept in its modern team sense. It is not the same as being comfortable or being nice; a psychologically safe team can hold a hard, direct conversation precisely because people trust that candour will not be used against them.

How do you measure psychological safety?

Measure it with a short, anonymous survey that people take on a regular cadence, so you get a trend rather than a one-off reading. Ask people to rate agree-or-disagree statements such as “I can raise a problem or a tough issue on this team” and “no one on this team would deliberately undermine my work”, then look at the spread and how the scores move over time. The point is to measure repeatedly and anonymously, because a single number tells you far less than a line that is rising or falling. A psychological safety health-check template runs this loop for you.

What are some psychological safety exercises?

Practical exercises include collaborative icebreakers that need everyone to take part, a facilitator openly asking the team for feedback on their own facilitation, a quick round-the-room check-in, and a team exercise where everyone requests and gives feedback. The most effective habit is small and repeated: model the behaviour you want by being the first to show vulnerability, deliberately pass the floor so the quieter people speak, offer anonymity as an option, and point to a concrete example where someone’s feedback led to a real change. Exercises help, but psychological safety is built by what the team does every week, not by a single workshop.

Who coined the term psychological safety?

Harvard Business School professor Amy Edmondson defined and popularised psychological safety in its modern, team-level sense in her research in the late 1990s. Her work reframed it from an individual trait into a property of a team: a shared belief, held by the group, that the team is safe for taking interpersonal risks like speaking up, disagreeing, and admitting error.

What did Google’s research find about psychological safety?

Google’s Project Aristotle studied its own teams to find out what separated the effective ones from the rest, and found that psychological safety was the single most important factor, ahead of dependability, structure and clarity, meaning, and impact. In other words, who was on a team mattered less than how the team worked together, and safety to take risks was the foundation the other dynamics were built on.