Psychological Safety at Work: What It Is and How Managers Build It

What is psychological safety at work?

In her 1999 paper in Administrative Science Quarterly, "Psychological Safety and Learning Behavior in Work Teams," Amy Edmondson defined team psychological safety as "a shared belief held by members of a team that the team is safe for interpersonal risk taking." Two words in that definition matter. Shared means it is a property of the team, not an individual personality trait; people on the same team tend to hold similar beliefs about it. Interpersonal risk means the risks are social: looking ignorant, incompetent, negative or disruptive in front of colleagues.

In everyday terms, a psychologically safe team is one where people say "I don't know," "I made a mistake," "I think this plan has a problem" or "Can you explain that again?" without first calculating the cost to their reputation. In a team without it, those sentences get swallowed, and the manager hears about problems late, if at all.

What did Amy Edmondson's research find?

Edmondson's 1999 study examined 51 work teams in a manufacturing company, measuring team conditions, team processes and outcomes. She found that team psychological safety was associated with learning behavior: seeking feedback, asking for help, discussing errors and experimenting. Team efficacy (the belief that the team is capable) was not associated with learning behavior once psychological safety was taken into account.

Learning behavior, in turn, mediated the relationship between psychological safety and team performance.

The practical reading is that psychological safety does not improve performance by making people feel good. It improves performance because teams that feel safe do the uncomfortable things that learning requires, like surfacing mistakes early and challenging assumptions. The paper also found that team structures, including context support and team leader coaching, shaped those shared beliefs.

In other words, leaders influence psychological safety directly.

Edmondson has continued the research since. A 2025 Harvard Business School Working Knowledge article described a study by Edmondson and Michaela Kerrissey, published in 2024, of more than 27,000 workers in a large U.S. hospital system surveyed in 2019 and 2021. In the 2021 data, a one standard deviation increase in psychological safety was associated with burnout lower by 0.72 points and willingness to stay higher by 0.63 points on a five-point scale, and the protective effect held even where staff reported inadequate tools or staffing.

What did Google's Project Aristotle find?

Google's People Analytics team studied 180 teams (115 engineering project teams and 65 sales pods), ranging from three to fifty members with a median of nine, to understand what made some teams more effective than others. According to Google's re:Work guide on team effectiveness, who was on a team mattered less than how the team worked together.

The researchers identified five dynamics that set successful teams apart, and ranked psychological safety first. The guide describes it as members feeling safe to take risks, such as admitting mistakes or offering new ideas, without fear of embarrassment or punishment. The other four are dependability, structure and clarity, meaning, and impact.

Equally useful is what the study found did not significantly predict effectiveness at Google: colocation of teammates, consensus-driven decision making, extroversion of team members, individual performance of team members, workload size, seniority, team size and tenure. Google is careful to note these factors may still matter elsewhere. But the finding is a useful corrective for managers who assume the answer to a struggling team is to hire stronger individuals.

What psychological safety is not

The term is often misread, which makes managers skeptical of it. Clearing up the misreadings makes it easier to act on.

• It is not niceness. A team can be polite and still unsafe, because people avoid conflict to keep the peace. Psychological safety makes candid disagreement easier, not rarer.
• It is not lower standards. Safety is about how people respond to risks and mistakes, not whether performance matters. The most effective teams combine high safety with high accountability.
• It is not comfort or the absence of criticism. Feedback can be direct and still safe, as long as it targets the work and the future rather than the person's worth.
• It is not a guarantee of agreement. Speaking up does not mean every idea gets adopted. It means ideas get heard and responded to honestly.
• It is not an individual trait. A confident person can feel unsafe on one team and safe on another. It describes the team climate.

How can you tell whether your team has psychological safety?

Most managers believe their team is safer than it is, because the evidence of an unsafe team is silence, and silence is easy to misread as agreement. Look for behaviors rather than asking people whether they feel safe.

• Who talks in meetings? If the same two or three people do most of the talking, others may be holding back.
• Do people report their own mistakes, or do you find out from customers or other teams?
• Does anyone ever disagree with you in a group setting?
• Do new joiners ask basic questions in public channels, or only in private messages?
• When a project goes wrong, does the conversation move to causes and fixes, or to who is to blame?
• Do people mention risks early in a project, or only when the deadline is already at risk?

Engagement surveys can add a quantitative view. Edmondson's original study used a short questionnaire with items such as whether mistakes are held against people and whether it is safe to take a risk on the team; many employee engagement platforms now include similar items. Gallup reported in 2017 that only three in 10 U.S. workers strongly agreed that at work their opinions seemed to count, and estimated that moving that ratio to six in 10 could be associated with a 27% reduction in turnover, a 40% reduction in safety incidents and a 12% increase in productivity.

Track the trend on your own team rather than comparing with benchmarks.

How do managers build psychological safety?

Psychological safety is built through hundreds of small interactions rather than a single workshop. The habits below fall into three groups that recur in Edmondson's work: how you frame the work, how you invite input, and how you respond when people take a risk.

1. Frame the work as uncertain. Say plainly that the project involves unknowns and that you expect problems to surface. People speak up more when raising issues is defined as part of the job rather than as complaining.
2. Admit your own limits. Saying "I'm not sure" or "I got that wrong last week" shows the team that fallibility is normal. The 2025 Working Knowledge piece describes Edmondson's advice to leaders to be open about not knowing.
3. Ask direct questions. "What am I missing?" and "Who sees this differently?" work better than "Any questions?" Call on quieter people with a specific, low-stakes prompt.
4. Make room before you give your view. In meetings, ask others to speak first. Once the most senior person states an opinion, many people will simply agree.
5. Thank the messenger. When someone brings bad news or admits a mistake, thank them before discussing the problem. The first five seconds of your reaction teach the team more than any policy.
6. Separate the person from the error. Run blameless reviews that ask what happened and what to change, not who failed.
7. Follow up on ideas. Close the loop on suggestions, including the ones you do not adopt, and explain why.
8. Hold standards openly. Be clear about what good work looks like. Safety without clarity tends to produce comfortable underperformance.

Edmondson's 2025 comments also stress building the habit before a crisis. People who have spoken up in difficult situations before and seen nothing bad happen are more likely to speak up again. A team that only hears "please raise concerns" in the middle of an emergency is unlikely to believe it.

Where do day-to-day routines fit in?

Routines are where psychological safety becomes visible. The weekly one-on-one meeting is often the first place someone tests whether it is safe to raise a problem, so how a manager handles the first small issue matters a lot. Retrospectives at the end of a sprint or project give the whole team a structured place to discuss what went wrong.

Human-centered project management makes these reviews a normal part of delivery rather than a reaction to failure.

Communication channels matter too. A team communication plan that sets norms for when to use public versus private channels, and an employee communication app that lets frontline staff raise issues without going through several layers of management, both lower the effort required to speak up. Anonymous feedback tools have a role, but if anonymity becomes the only way people will raise issues, that is a sign the team climate needs work.

Finally, specific and fair employee recognition reinforces the behaviors you want. Publicly thanking someone for flagging a risk early tells everyone else that flagging risks is valued.

What gets in the way?

Several common management habits quietly erode psychological safety, often without the manager noticing. Interrupting, visibly checking a phone while someone speaks, reacting with irritation to questions, praising only the people who agree, and discussing a person's mistake with others before discussing it with them all send the same signal.

Structural pressures matter as well. Teams under constant deadline pressure, frequent reorganizations or layoffs often see psychological safety drop, because people become more careful about how they appear. The 2024 hospital study described above suggests that safety can act as a buffer in difficult conditions, but it does not remove the need to fix the conditions themselves.

Chronic overload is a separate problem, covered in our guide to preventing team burnout.

Remote and hybrid work changes the signals rather than removing them. On video calls it is harder to read hesitation, so managers need to ask more explicitly and use written channels where quieter people may feel more comfortable contributing.

Frequently asked questions

Who coined the term psychological safety?

Edgar Schein and Warren Bennis discussed psychological safety in 1965 in the context of organizational change, but the team-level concept used today comes from Amy Edmondson of Harvard Business School. Her 1999 paper in Administrative Science Quarterly defined team psychological safety as a shared belief that the team is safe for interpersonal risk taking, and tested it in a study of 51 work teams in a manufacturing company.

What was Google's Project Aristotle?

Project Aristotle was a Google People Analytics study of 180 internal teams that looked at whether team composition or team dynamics best explained effectiveness. Google's re:Work guide reports five dynamics of successful teams: psychological safety, dependability, structure and clarity, meaning, and impact. Psychological safety ranked first, and factors such as team size, seniority and colocation were not significantly connected to effectiveness at Google.

Does psychological safety mean nobody can be criticized?

No. Psychological safety makes it easier to give and receive candid criticism, because people trust that feedback is about improving the work rather than attacking them. Teams that avoid all criticism to keep the peace are often less safe, since people fear the consequences of breaking the polite surface. The goal is honest, respectful disagreement combined with clear performance standards.

How do you measure psychological safety?

Short survey scales are the most common method. Edmondson's original research used a questionnaire with statements about whether mistakes are held against people and whether it is safe to take risks on the team. Many engagement survey tools include similar items.

Combine survey results with observation: who speaks in meetings, whether people report their own mistakes, and how the team discusses failures.

Can a single manager build psychological safety if the wider company culture is poor?

Yes, to a meaningful degree. Edmondson's research treats psychological safety as a team-level belief, and her 1999 study found that it varied across teams within the same company. A manager who responds well to bad news, admits mistakes and protects people who speak up can create a safer climate on their own team, although company-wide pressures such as layoffs will still have an effect.

How long does it take to build psychological safety?

There is no fixed timeline, because it depends on the team's history and how consistently the manager behaves. Trust builds through repeated experiences of speaking up without negative consequences, so progress is gradual. It can also be lost quickly: a single harsh public reaction to a mistake can undo months of work, which is why consistency matters more than any single initiative.

Sources and further reading

• Psychological Safety and Learning Behavior in Work Teams (Amy Edmondson, Administrative Science Quarterly, 1999), Administrative Science Quarterly / Cornell University
• Guide: Understand team effectiveness, Google re:Work
• In Tough Times, Psychological Safety Is an Asset, Not a Luxury, Harvard Business School Working Knowledge
• How to Create a Culture of Psychological Safety, Gallup