In the mid-1990s, Amy Edmondson began studying medication errors in hospital units. She expected the better-run teams to make fewer of them.
The data said the opposite. Units with stronger relationships and better manager coaching reported more errors.
Read literally, good teamwork was making patients less safe. That made little sense. The better teams might not be making more mistakes at all. They might simply be more willing to admit them.
That is the useful heart of psychological safety. It is not a promise that work will be pleasant, nobody will argue, or every idea will receive a warm round of applause. It is whether people will say what they know when saying it could make them look ignorant, difficult, or wrong.
Confuse safety with niceness and a team can become wonderfully polite while the important information moves to the hallway. The research points away from slogans and toward something more mundane: what happens in the room immediately after someone says the plan has a hole in it.
Key takeaways
- Psychological safety means a team can admit mistakes, ask questions, and disagree without expecting punishment or humiliation.
- Amy Edmondson's hospital research found that stronger teams surfaced more errors. Silence made weaker teams look safer than they were.
- Psychological safety and high standards are separate. A team needs both.
- Current workplace data links lower psychological safety with toxic workplace experiences.
- Managers build safety through their reactions to bad news, not through posters, slogans, or a newly renamed meeting.
Why did the better hospital teams report more errors?
Edmondson's 1996 hospital study examined medication errors across 8 patient-care units in 2 urban hospitals. It included responses from 123 nurses, 19 physicians, and 4 pharmacists.
The paper found systematic differences not only in reported errors, but in whether teams detected, discussed, and learned from them. An error count was also a count of what people were willing and able to surface.
That distinction matters. The study did not prove that every unit made the same number of mistakes, or that strong teams were secretly flawless. Actual error rates were unknowable. What it showed was that a quiet unit could look clean because problems stayed hidden.
The quiet teams were not necessarily safer. They were quieter.
What psychological safety actually means
Psychological safety is a shared belief that a team is safe for interpersonal risk-taking. People can raise concerns, ask for help, admit mistakes, or challenge an idea without expecting embarrassment, rejection, or punishment. It describes a team climate, not a personality trait, a promise of comfort, or permission to lower standards.
Edmondson gave the concept its modern team-level definition in her 1999 study of 51 work teams. The study connected psychological safety with learning behavior: seeking feedback, discussing errors, asking for help, and trying new approaches.
The definition is about consequences. If I say what I actually think, what happens next?
On a psychologically safe team, the answer is not necessarily agreement. The team deals with the information rather than punishing the person who supplied it.
Niceness and trust are not psychological safety
Niceness often avoids a difficult conversation to preserve the mood. Psychological safety lets the conversation happen.
Edmondson and Harvard public-health professor Michaela Kerrissey opened their 2025 review of 6 common misconceptions with this exact confusion. Politeness becomes a liability when it stops a team from sharing information the work needs.
A merely nice team says the plan looks fine, then explains in the hallway why it will fail. A psychologically safe team says it in the meeting, while there is still time to fix it.
That does not make rudeness useful. Candor still needs respect, evidence, and judgment. "I am just being honest" has covered a great deal of ordinary bad behavior. Psychological safety protects useful interpersonal risk; it does not grant diplomatic immunity to whoever speaks first or loudest.
Trust is usually about a relationship between people. Psychological safety is a shared belief about the team: what happens here when someone asks for help, raises a concern, or makes a mistake?
The two overlap, but they are not interchangeable. You may trust a colleague personally and still stay quiet in a meeting because the manager punishes dissent. You may also work on a psychologically safe team without feeling close to everyone on it. The useful question is not whether the team likes one another. It is whether important information can travel upward, sideways, and into the room.
Psychological safety still needs high standards
Safety is not the problem. Safety without standards is.
Edmondson treats psychological safety and performance standards as separate axes. In a Harvard Business Review conversation about blame and accountability, she described the resulting 4 zones and why "safe" and "demanding" are not opposites:
| Low standards | High standards | |
|---|---|---|
| Low safety | Apathy | Anxiety |
| High safety | Comfort | Learning and performance |
Low safety with high standards creates anxiety. People know the bar is high, so they hide problems that might make them look incapable. High safety with low standards creates comfort, which can be pleasant and unproductive in equal measure.
The goal is the final quadrant: people can report a miss, and the miss still matters. A deadline does not disappear because someone admitted it was in danger. The admission gives the team a chance to do something before the deadline becomes an archaeological discovery.
What current evidence can and cannot tell us
Psychological safety is associated with how workers experience their jobs, although a survey cannot prove that it causes every outcome alongside it.
The American Psychological Association's 2024 Work in America survey covered 2,027 employed adults. Overall, 15% described their workplace as somewhat or very toxic. Among that group, 89% also reported lower psychological safety.
The same survey found that workers with higher psychological safety were more likely to report satisfaction with coworkers, managers, schedules, growth opportunities, and employer support for mental health. The pattern is not mysterious. Work is easier to improve when the people doing it can describe what is wrong.
The underlying research base is substantial, but the label is now used more loosely than the original construct. Edmondson's 1999 study tested a specific team-level belief and a specific set of learning behaviors. The phrase now appears in discussions of wellbeing, civility, inclusion, accountability, leadership style, and general workplace happiness. Those subjects overlap. They are not all the same variable wearing a conference badge.
That is why the definition matters. Psychological safety does not promise that work will feel easy, that every idea will be adopted, or that nobody will face consequences. It means people can contribute relevant information without interpersonal punishment for the act of speaking.
Keep the concept that narrow and it remains useful.
Managers teach safety one reaction at a time
Psychological safety can vary sharply from team to team inside the same organization because it is built through local behavior.
The company may have a polished values page. The team watches what the manager does when someone says the launch is at risk. One reaction carries more instructional value than a quarter's worth of culture programming.
That is good news for managers. You do not need a company-wide transformation program before your own team can improve. You need to react well the next time someone brings you information you did not want.
Managers build psychological safety by making candor useful and survivable. Four habits matter more than a grand announcement.
React to mistakes with questions before judgment. Start with what happened, what the team knows, and what needs containing. Accountability still comes, but a public hunt for the guilty teaches everyone else to improve their hiding technique.
Ask for dissent before the decision. "Any questions?" usually produces the silence it deserves. Ask what could make the plan fail, what assumption feels weakest, or what a skeptical customer would notice first.
Show your own uncertainty accurately. Pretending to know everything makes other people hide what they do not know. Inventing fake vulnerability is worse. Name the real uncertainty, then ask who has information that changes it.
Reward early bad news. Problems get more expensive with age. Thank the person who surfaced one while it was still small enough to be boring.
What an org chart can show
An org chart shows formal structure: titles, reporting lines, and who is accountable to whom. That clarity matters, particularly when a team is growing or reorganizing. Our guide to creating an organizational chart covers the structural part.
The chart cannot show whether the person 3 boxes down feels safe telling the person at the top that the plan has a hole in it. It cannot render trust, candor, or the consequences of speaking up.
That was also the central limit in Google's Project Aristotle. Google's research identified psychological safety as the strongest of 5 team dynamics, while the org chart could display only part of one: structure and clarity.
Draw the reporting lines clearly. Then remember that the important information still has to cross them.
Frequently asked questions
What is psychological safety at work?
Psychological safety at work is a shared team belief that people can ask questions, admit mistakes, raise concerns, and disagree without being humiliated or punished. It concerns the interpersonal consequences of speaking up. It does not guarantee agreement, comfort, promotion, or protection from ordinary performance standards.
Is psychological safety the same as being nice?
No. Niceness can mean avoiding conflict to keep a conversation comfortable. Psychological safety makes respectful conflict possible. A safe team can challenge an idea, deliver difficult feedback, and admit a failure while still treating the people involved with dignity.
Who developed the idea of psychological safety?
The term existed earlier, but Harvard Business School professor Amy Edmondson established the modern team-level construct through her 1999 research. Her earlier hospital work showed why it mattered: better-functioning units often detected and discussed more errors, revealing how silence can distort performance data.
Can psychological safety reduce accountability?
It can be misused that way, but reduced accountability is not part of the concept. Psychological safety and high standards are separate conditions. Strong teams make it safe to report a problem and still expect people to help solve it, learn from it, and meet the agreed standard.
How can a manager improve psychological safety?
Respond to bad news with questions before blame, ask for specific dissent before decisions, admit genuine uncertainty, and reward people who surface problems early. Employees learn whether speaking up is safe from a manager's repeated reactions, not from a policy statement.
Related guides
- What Google's Project Aristotle found about effective teams
- Why team dependability quietly breaks down
- How to create an organizational chart
An org chart could have shown who worked in each hospital unit. It could not have shown which team would admit an error.
The chart shows who is in the room. It cannot make anyone speak up in it.