# Psychological safety: what the research says and what managers can do Source: https://sarang.pro/psychological-safety-research-managers/ Author: Sarang | Published: 25 October 2026 Most managers say they want to hear bad news early. Fewer have teams that bring it to them. The gap usually has little to do with the size of the problem and a lot to do with what people expect will happen to them if they raise it. Researchers call the missing ingredient psychological safety. The term is now everywhere, which has blurred what it means. This article sets out what the research says, what the idea is often mistaken for, and the everyday behaviours that build it. ## Amy Edmondson’s definition Amy Edmondson, a professor at Harvard Business School, is the researcher most closely linked with team psychological safety. Her definition, as quoted by Google’s re:Work, is “a shared belief held by members of a team that the team is safe for interpersonal risk taking.” Two words in that sentence matter. “Shared” means it belongs to a team, not to one confident individual. “Interpersonal risk” means the small social risks of work: admitting a mistake, asking a question that might sound naive, disagreeing with a senior colleague or suggesting an unusual idea. Her best-known paper, “Psychological Safety and Learning Behavior in Work Teams”, appeared in *Administrative Science Quarterly* in 1999. Earlier hospital research, published in 1996, pointed in the same direction. Units where staff believed they would not be punished for reporting mistakes detected more errors. Those units talked about errors more openly, which made people more willing to report them in future. The lesson for any manager: a team that reports few problems may simply be a team that hides them. ## What Google’s Project Aristotle found Google studied 180 of its own teams, 115 engineering project teams and 65 sales pods, to find out what made some more effective than others. According to Google’s re:Work guide, five dynamics stood out. In order of importance: - **Psychological safety:** people feel able to take interpersonal risks without fear of being mocked or penalised. - **Dependability:** teammates get good work done on time. - **Structure and clarity:** people know what is expected of them, how to meet it and what their work will lead to. - **Meaning:** the work matters personally to the people doing it. - **Impact:** people believe their work makes a difference to the organisation’s aims. Psychological safety was the strongest of the five. The researchers concluded that how a team worked together mattered more than who was on it. Several factors showed no significant link to effectiveness at Google, including colocation, consensus decision making, extroversion, individual performance, workload size, seniority, team size and tenure. Google notes that these could still matter in other organisations, so read the list as findings about Google rather than universal laws. ## What psychological safety is not In a 2025 *Harvard Business Review* article, Edmondson and Michaela Kerrissey set out six misconceptions that have led organisations astray. In short, psychological safety does not mean: - **Being nice.** Politeness can hide disagreement. A team where everyone is pleasant and nobody challenges a weak plan may feel comfortable, but nobody in it is taking interpersonal risks. - **Getting your way.** People need to be heard. A manager can listen carefully and still decide against a suggestion. - **Job security.** It means freedom to speak candidly and makes no promise about employment. - **A trade-off against performance.** Safety and accountability are separate dimensions. High safety with low standards does not produce excellent work. - **A policy.** It is built in everyday conversations and cannot be created by a rule. - **Something that must come from the top.** Safety varies from team to team, so any manager can build it in their own team. Edmondson’s book *The Fearless Organization* (2018) adds a related point. Responding fairly and in proportion to clearly blameworthy conduct, such as bullying or dishonesty, reinforces psychological safety. People feel safer when they can see that the boundaries are real. ## Safety and standards belong together The most common objection from managers is that a “safe” team will become a soft one. The research points the other way. Safety lets people say “I am behind”, “I made a mistake” or “I think this plan is wrong” early enough to act on it. Standards make sure the work still has to be good. In uncertain work, where nobody knows the answer in advance, the Edmondson and Kerrissey summary suggests both need to be high. An example (illustration only): a small e-commerce team misses a product feed deadline. In a low-safety team, the person responsible stays quiet until the listings fail. In a high-safety, high-standards team, they flag the risk two days early, the manager thanks them, and the team agrees what to cut so the deadline holds. The deadline matters in both teams. The second one simply hears about the problem in time. ## What managers can do Edmondson’s leader’s toolkit has three parts: - **Set the stage.** Frame the work and its stakes. Say out loud when a project is uncertain and that you will need people’s input to get it right. - **Invite participation.** Make specific, structured requests for input. “My door is always open” is not an invitation; “What is the biggest risk you can see in this plan?” is. - **Respond productively.** Thank people for raising problems, treat failure as information and deal firmly with clear violations. The practical advice in the 2025 HBR article points the same way. Model the behaviour by asking questions you do not know the answer to and admitting when you do not know something. Watch your meetings: are people contributing, listening and pairing their own views with questions? Build regular reflection habits, such as a weekly round of progress updates, so problems surface before a big reveal. Some small, practical habits that put this into action: - Speak last in discussions where your opinion might close down others. - End decisions with “What are we missing?” and wait through the silence. - When someone brings bad news, thank them before you ask what happened. - Admit your own mistakes plainly and say what you will do differently. - Review failures by asking what in the process allowed them, not who is to blame. ## Limits worth knowing Psychological safety is not a cure-all. The research summary on Wikipedia notes studies with less comfortable findings: one found that psychological safety strengthened the link between utilitarian-minded team members and unethical behaviour such as cheating, and another linked higher psychological safety with lower work motivation. These are reminders that safety works alongside clear goals and standards, not in place of them. It also cannot be declared. A manager who announces that the team is now psychologically safe has changed nothing. What changes the climate is the next time someone raises a problem, and what happens to them when they do.