Psychological Safety: How to Build It on Your Team

Psychological safety is not a feel-good policy. Psychological safety is a shared belief within a team that it is safe to ask questions, admit mistakes and raise concerns without being embarrassed or punished. Amy Edmondson of Harvard Business School defined it that way in 1999. Anyone who works on a team like that speaks up sooner when something is going wrong. That is exactly why less goes wrong there.

This guide follows Priya, 36, an engineering manager at a health insurance software company in Denver. She leads six engineers who ship the claims portal. Her team looked healthy on paper: friendly standups, few complaints, on-time releases. Then a config change took the portal down for four hours on a Monday, and the postmortem revealed that a junior engineer had spotted the risk two weeks earlier and said nothing. Along the way you get Edmondson's definition, the 4 stages of psychological safety, seven survey questions, six steps you can copy and a plan from yearly goal to daily standup.

What Is Psychological Safety? Edmondson's Definition

Edmondson describes psychological safety as a shared belief that the team is safe for interpersonal risk taking. Interpersonal risks are small moments with big consequences: a question that might sound naive, a mistake you would rather hide, a disagreement with your boss. Anyone who expects those moments to be used against them stays quiet. That means: psychological safety is not one person's feeling. It is a climate the whole team shares.

The idea is older than Edmondson's study. Edgar Schein and Warren Bennis used the term in 1965 in their work on organizational change. William Kahn applied it to individuals in a 1990 paper in the Academy of Management Journal: you are psychologically safe when you can bring your full self to work without fearing damage to your image, status or career. Edmondson then lifted the concept to the team level and gave it a way to be measured.

Her 1999 study looked at 51 work teams in a manufacturing company. Teams with higher psychological safety showed more learning behavior. They asked for feedback, discussed errors and tried new approaches more often, and that learning behavior was linked to better performance. In short: for Edmondson, safety is not the goal in itself. It is what allows a team to learn from what goes wrong.

What Psychological Safety Is Not: Three Myths

The phrase gets used a lot and misunderstood almost as often. Three myths come up in nearly every leadership workshop. They sound reasonable. They just lead in the wrong direction:

• “Psychological safety means being nice.” Kindness helps. It can also mean nobody pushes back because no one wants to spoil the mood. Edmondson is clear that psychological safety exists to make candor possible.
• “Psychological safety lowers the bar.” The opposite is true. Safety and performance standards are two separate dimensions. A team needs both to learn and deliver.
• “Psychological safety is the same as trust.” Related, but not identical. Trust is about how you see one specific person. Psychological safety is about how you expect the group to treat you when you take a risk.

Priya had mistaken the first myth for a strength. Her standups were warm, and conflict was rare. What she missed was the price of that calm. Anyone who wanted to raise something uncomfortable waited for a better moment, and the better moment never came. Is a team where nobody disagrees really safe? Usually it is just quiet.

A fourth myth concerns accountability. Psychological safety does not mean mistakes have no consequences. Anyone who knowingly ignores a security policy still faces the result. The difference is elsewhere. A psychologically safe team never punishes someone for reporting a mistake. If anything, it treats hiding one as the real problem.

Why Team Psychological Safety Drives Performance

The best-known evidence comes from Google. Starting in 2012, its Project Aristotle studied more than 180 teams to find out what made some of them effective. In 2015 Google reported that who was on a team mattered less than how the team worked together. Of five factors, psychological safety came first, ahead of dependability, structure and clarity, meaning and impact.

Edmondson's own path to the topic began with a puzzle. In a hospital study published in 1996, the better-led nursing units reported more medication errors, not fewer. The explanation was that they did not make more errors. They reported more of them. Anyone who does not fear blame also reports the mistake nobody else would have noticed.

Priya found the same pattern in her own numbers. Her team had filed only two incident reports in the previous quarter, which had looked like a sign of quality. For a team shipping weekly to 40,000 users, it was more likely a sign of silence. That means: a low error count on paper can also mean that nobody is writing errors down.

A 2017 meta-analysis by M. Lance Frazier and colleagues in Personnel Psychology pulled the research together. Across the studies it reviewed, psychological safety was linked to more information sharing, more learning behavior and better performance. It is still not a cure-all. It does not replace clear goals, training or good planning. What it does is make sure a team uses the information it already has. The idea is old: in Maslow's hierarchy of needs, safety comes right after the physical basics, and a person who feels threatened spends energy on protection instead of curiosity.

The 4 Stages of Psychological Safety

Timothy R. Clark described the 4 stages of psychological safety in a 2020 book. The stages build on each other, so a person rarely reaches the fourth without the first three. They are useful when you want to understand where one individual stands, not just the team on average.

The junior engineer who spotted the risky flag was stuck at stage two. She was still learning the system and did not want to look like she was guessing. Challenging a senior colleague's config change would have required stage four. That gap between where she was and where the situation needed her to be is where the outage came from.

Safety and standards: the four zones

Edmondson adds a second lens. She sorts teams by two questions: How safe do people feel? And how high are the performance standards? The answers produce four zones:

Priya rated her team in the comfort zone. Her predecessor had left behind the anxiety zone, with tight deadlines and public blame after every bug. Both lead to the same result. The most important information never makes it into the room. Anyone who works on only one axis moves the team into the neighboring zone, not into the learning zone.

Ad

Prosp

Prosp · LinkedIn outreach

LinkedIn outreach that gets repliesPersonal messages plus AI voice notes in your own voiceEvery reply in one inbox

Try for free→

New reply

Hi Anna, your post on remote sales got me curious.

Sounds interesting. Does Thursday work?

How to Measure Psychological Safety: Seven Questions

Team psychological safety can be measured, and it takes very little effort. Edmondson's 1999 study used seven statements, each rated on a scale from 1 to 7. Three are negatively worded and scored in reverse. Paraphrased, they read like this:

Priya ran the seven statements as an anonymous form with no name field and no team breakdown. Seven of eight people answered, including the two contractors. The average came out at 4.3 out of 7. The weakest item was number one, mistakes held against you, at 3.4. The strongest was number six at 5.9. People did not suspect each other of bad intent. They simply remembered the old blame culture.

Three tips for your own survey. First, anonymous has to mean anonymous, so skip any breakdown that would identify people on a small team. Second, the number is a starting point, not a grade. Third, repeat the survey in the same format after three months, or you will never know whether anything changed. Anyone who runs a survey and never discusses the results with the team has turned a measurement into another disappointment.

How to Build Psychological Safety in the Workplace: Six Steps

In “The Fearless Organization”, Edmondson describes three jobs for leaders: set the stage, invite participation and respond productively. Priya turned them into six steps for her team. None of them takes long. Together they change what gets said in standup.

1. 01

   Frame the work, 10 minutes

   Tell the team out loud why every signal matters. Priya said in the next retro: “We deploy to 40,000 people every week. At that pace, things will break. I want to hear about every risk before a customer does.”

2. 02

   Show your own uncertainty, 2 minutes

   Admit what you do not know. Priya told the team she had approved the config change herself without reading the flag description. A leader who admits mistakes lowers the cost of admitting them for everyone else.

3. 03

   Ask specific questions, 5 minutes a day

   Almost nobody answers “Any questions?” A specific question works better: “What almost bit us yesterday?” Priya asked it at the end of every standup and then waited a full ten seconds.

4. 04

   Respond well to the first report, right away

   The first person who speaks up is testing the waters for everyone. Thank them specifically, ask for details and say what happens next. When the junior engineer flagged a second risky flag, Priya pinned her message in the team channel with a thank-you.

5. 05

   Run blameless reviews, 30 minutes a week

   Set a fixed slot to review incidents and near misses without looking for someone to blame. The guiding question is: What about our process made this mistake easy? Google's 2016 book on site reliability engineering describes the same idea as a blameless postmortem culture.

6. 06

   Close the loop visibly, 5 minutes a week

   Nothing kills candor faster than reports that disappear. Priya kept a shared list of every flagged risk with an owner and a date. After six weeks, eight of eleven items were marked done.

Step four matters most and breaks most easily. Edmondson calls it responding productively: express appreciation, destigmatize failure and sanction only clear violations. One irritated sentence can undo weeks of work. Anyone who greets the first uncomfortable report with a sigh will not get a second one.

Many of these conversations happen one on one, not in the group. The one-on-one meeting questions are a good place to start. When you need to give critical feedback without damaging safety, follow the structure in how to give feedback to employees: describe the behavior, name the impact and agree on what happens next.

Warning Signs That Safety Is Slipping

Psychological safety rarely disappears with a bang. It erodes quietly. Priya wrote down these signals so she would notice early if her team started sliding back:

• Meetings get shorter. What looks like efficiency often means nobody is contributing.
• Problems surface late. Anyone who learns on Thursday about a bug found on Monday has a reporting problem, not a bug problem.
• The same two voices. When only the most senior people talk in retros, the others are not short on ideas.
• Jokes about past mistakes. A nickname for the person who once broke production sounds harmless. It is a warning to everyone else.
• Feedback only in private. When criticism reaches you only in direct messages, nobody feels safe saying it in front of the team.

Priya walked it back the next day. She told the team her response had been wrong and asked the engineer to explain the concern again. It took two minutes. Two new risks came up in the same meeting. Mistakes on the leadership side are part of learning too, as long as they are corrected in the open.

Psychological Safety in the Workplace: What Leaders Can and Cannot Do

Psychological safety in the workplace is built mostly inside the immediate team. Company culture sets the frame, but whether it is safe to report a mistake is something each person learns from their direct manager. That is good news. You do not need to wait for a company-wide culture program to change things on your team.

Leadership also has limits. Psychological safety is a team climate, not therapy and not a substitute for health support. The U.S. Surgeon General's 2022 Framework for Workplace Mental Health and Well-Being lists psychological safety as part of protection from harm, next to things like workload and fair treatment. Those are related topics with their own tools, and a manager is not the right person to handle all of them alone.

Anyone who notices a team member under heavy strain for a long time should take it seriously and point them to an employee assistance program, HR or a professional rather than trying to treat it themselves. If you or someone on your team is in crisis, call or text 988 to reach the Suicide and Crisis Lifeline in the US.

What about poor psychological safety? It rarely looks dramatic. It looks like silence in meetings, surprises at the end of a sprint and good people leaving without saying why. Improving it does not start with perks or offsites. It starts with how the next reported mistake is handled.

From Yearly Goal to Daily Standup

Psychological safety sounds like something you cannot plan. That is only half true. You cannot order a climate into existence. You can plan the habits that create it. Priya broke her plan down across all five levels like any other goal.

The yearly goal has two numbers on purpose. The survey score shows how safe people feel. The report count shows whether that feeling changes behavior. Anyone who measures only the feeling can confuse a good mood with the learning zone. Anyone who counts only reports risks people filing trivial ones to hit a quota.

Psychological safety is also never one person's goal. It grows where a team shares goals and talks honestly about the path there. The guide to team goals shows how to set those together. The article on goals for managers shows how to write your own version as a leadership goal. In Gaapp, Priya entered the yearly goal once and now sees every Friday whether the week's rituals actually happened.

After three months her team scored 4.9. Near-miss reports went from two in the previous quarter to nine. The yearly goal was not reached yet. But the direction was right, and for the first time it was visible.

Self-Check: Three Questions for Your Friday

Psychological safety shifts with every reaction you show. A short look at the end of each week is worth it. Anyone who cannot answer one of these questions has found their next step:

• When did someone tell me something this week that I did not want to hear, and how did I respond?
• Who on the team said nothing at all this week, and what might be behind that?
• Which flagged risk from the last few weeks is still open, and does the person who raised it know why?

Priya answers the three questions every Friday afternoon in five minutes and writes the answers next to her weekly goals. After a quarter she can see not only whether the rituals happened but also how she reacted. In short: psychological safety is not a state a team reaches. It is a promise a leader keeps again every day.