Guide · Team development · HR

Psychological safety in teams: the research, the models and what actually works

Why do smart people agree to decisions nobody wants? Why do some groups outperform the sum of their members? This guide walks through 70 years of research – from the Johari Window and the Abilene Paradox to Google's Project Aristotle, Wheelan's IMGD and Sweden's UGL – and shows how to put it to work in everyday teamwork.

A team in a sunlit office during a collaborative meeting
Photo: Unsplash

Key takeaways

  • Psychological safety is a team's shared belief that it is safe to take interpersonal risks: to ask, to admit mistakes, to disagree.1
  • It was the most important factor for effective teams in Google's study of 180+ teams5 and is linked to learning, engagement and performance in a meta-analysis of ~22,000 people.4
  • The Johari Window explains the mechanism: safety makes us willing to give feedback and reveal more of ourselves, so the "open window" grows.9
  • The Abilene Paradox shows what happens without it: a group agrees on something nobody actually wants.15
  • Groups mature in stages (Tuckman, Wheelan), and safety can be trained – through programmes like UGL, measures like GDQ and small everyday habits.

What is psychological safety?

In 1999 Harvard professor 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.1 It is not about everyone agreeing or being nice. It is about being able to say "I don't know", "I made a mistake" or "I think we're heading in the wrong direction" without paying a social price.

Edmondson's interest began with a surprising finding. In a study of hospital teams, the units with the best leadership and relationships reported more medication errors – not because they made more mistakes, but because they were willing to talk about them.2 In her follow-up study of 51 work teams in a manufacturing company, psychological safety predicted team learning behaviour, which in turn predicted performance.1

The concept builds on William Kahn's work on engagement, which identified psychological safety as one of three conditions for people to "bring themselves" fully to their work.3

Psychological safety isn't about lowering standards. It's about creating a climate where high standards can be met – because problems surface in time. Paraphrasing Amy Edmondson, The Fearless Organization (2019)6

What psychological safety is not

  • Not conflict avoidance. Safe teams often argue more about the work – precisely because the relationships can take it.
  • Not low standards. Edmondson calls the combination of high safety and high standards the "learning zone"; safety without standards becomes a "comfort zone".6
  • Not a personality trait. Safety lives in the group and its norms, not in the individual. The same person can be outspoken in one team and silent in another.

What the research says: Project Aristotle and collective intelligence

In 2012 Google launched Project Aristotle to find out why some teams succeed and others don't. Researchers studied more than 180 teams and tested hundreds of variables – education, personality, gender, whether people sat in the same office and much more. The conclusion: who is on the team matters less than how team members interact.57

Five dynamics stood out, and psychological safety was by far the most important – the one that makes the other four possible:

  1. Psychological safety – can we take risks without feeling insecure or embarrassed?
  2. Dependability – can we count on each other to deliver quality work on time?
  3. Structure and clarity – are roles, goals and plans clear?
  4. Meaning – is the work personally important to us?
  5. Impact – do we believe our work makes a difference?

These findings fit closely with research on collective intelligence. In 2010 Anita Woolley and colleagues at MIT and Carnegie Mellon showed in Science that groups have a measurable "c factor" predicting how well they solve very different tasks. It was not strongly linked to members' average or maximum IQ – but it was linked to social sensitivity (the ability to read others) and to how evenly speaking time was shared.8 Alex Pentland's measurements of communication patterns point the same way: in high-performing teams, everyone talks and listens in roughly equal measure.29

180+teams were studied in Google's Project Aristotle5
136independent samples (~22,000 people) in Frazier's meta-analysis4
10–15%are truly self-aware – yet around 95% believe they are14

The largest synthesis to date, a meta-analysis by Frazier and colleagues (2017), covers 136 independent samples with more than 22,000 individuals and almost 5,000 groups. Psychological safety was clearly associated with information sharing, learning, engagement and performance – for individuals and for teams.4

A laptop showing a video meeting with several participants
Project Aristotle found that sitting in the same office was not what made the difference – interaction matters more than location. Photo: Unsplash

The Johari Window: the model behind self-awareness and feedback

Back in 1955, psychologists Joseph Luft and Harrington Ingham at UCLA introduced a simple model of how we are perceived by others. The name "Johari" combines their first names.910 The model divides what is known about a person into four windows, depending on whether the information is known to the person and to others:

The Johari Window (after Luft & Ingham, 1955). The goal is a growing open window – by giving and receiving feedback and by revealing more of ourselves.

The point of the model is that a larger open window makes collaboration easier: fewer misreadings, less energy spent guessing and more trust. The open window grows in two ways – when others tell us how they perceive us (the blind spot shrinks) and when we tell others more about ourselves (the façade shrinks). Both require psychological safety.

Modern research confirms the blind spots

Johari is a teaching model, but later research has given it solid empirical support:

  • Others see us partly better than we see ourselves. Simine Vazire's SOKA model shows that we are best at judging "internal" traits such as anxiety, while others are often better at judging evaluative traits such as intelligence and creativity.11
  • We think we know others better than they know us. Pronin and colleagues call this the illusion of asymmetric insight.12
  • We overestimate how transparent we are. Under the illusion of transparency, we believe others can read our feelings and intentions more clearly than they actually can.13
  • Self-awareness is rarer than we think. Tasha Eurich's research suggests that around 95% of people believe they are self-aware, while only 10–15% actually are by her criteria.14

Self-disclosure also works as social glue. A meta-analysis by Collins and Miller shows that we like people who open up, and that we open up more to people we like.27 In Arthur Aron's classic experiment, 45 minutes of gradually more personal questions was enough to create real closeness between strangers.28

The Abilene Paradox: when everyone agrees to something nobody wants

On a hot summer afternoon in Coleman, Texas, a father-in-law suggests the family drive to Abilene for dinner – about 50 miles away. His daughter says it sounds like a great idea. Management scholar Jerry B. Harvey agrees, because he doesn't want to spoil things. His mother-in-law says she would love to come. Four hours later, after a dusty trip without air conditioning and a mediocre meal, it turns out that none of them wanted to go. Everyone said yes because they thought the others wanted to.15

In 1974 Harvey described this as the Abilene Paradox: organisations often fail not because they can't manage conflict, but because they can't manage agreement. Each individual knows what they think – but nobody says it, for fear of being left out.

A person presenting ideas at a whiteboard to colleagues
"Is everyone on board with this?" Silence rarely means everyone agrees. Photo: Unsplash

Abilene, groupthink and false consensus

The Abilene Paradox is often confused with groupthink, which Irving Janis described in 1972 using cases such as the Bay of Pigs invasion.16 The difference matters:

PhenomenonWhat happens?How to break it
Groupthink (Janis 1972)The group believes in the decision and suppresses dissenting information to preserve cohesion.A devil's advocate, outside reviewers, the leader speaking last.
Abilene Paradox (Harvey 1974)Most members oppose the decision, but each assumes the others are for it.Surface what each person actually thinks – ideally simultaneously and independently.
False consensus (Ross et al. 1977)We overestimate how many people share our own opinions and choices.Ask instead of assuming. Compare guesses with real answers.
Organisational silence (Morrison & Milliken 2000)Employees withhold concerns and ideas because they believe speaking up is pointless or dangerous.Leaders who invite and reward uncomfortable information.

In Lee Ross's classic study of the false consensus effect, students were asked whether they would walk around campus wearing a sandwich-board sign. Both those who agreed and those who refused believed that most other people would make the same choice as they did.17 This is exactly the mechanism that drives the family to Abilene – and leadership teams to decisions nobody really supports. Morrison and Milliken showed how it can become an entire organisational culture of silence.18

The Abilene test for your next meeting

Before an important decision, ask everyone to write down their own position (1–5) and their guess about what the group thinks – before anyone speaks. Reveal the answers at the same time. The gap between guesses and reality is often an eye-opener.

How groups mature: Tuckman, Wheelan (IMGD) and FIRO

Psychological safety doesn't appear overnight. Groups move through typical stages, and what works in one stage can be wrong in another.

In 1965 Bruce Tuckman summarised the research in the familiar sequence forming, storming, norming, performing, adding adjourning with Mary Ann Jensen in 1977.19 The model was later developed and empirically tested by Susan Wheelan, whose Integrated Model of Group Development (IMGD) is widely used by organisations in Scandinavia and beyond.2021

  1. Dependency & inclusionMembers are polite, look to the leader for direction and avoid conflict.
  2. Counter-dependency & fightDifferences in values and ways of working surface. Necessary – but many teams get stuck here.
  3. Trust & structureThe group negotiates roles and norms. Safety grows.
  4. Work & productivityFocus on goals and results. Feedback is given and received openly.

Wheelan also developed the GDQ (Group Development Questionnaire), which shows which stage a group is in. Her studies found, among other things, that groups in later stages perform better – and that getting there takes time, often many months.20 Today GDQ is managed and developed by GDQ Associates, in collaboration with researchers at the University of Gothenburg.22

FIRO – the needs beneath the surface

Will Schutz's FIRO theory (Fundamental Interpersonal Relations Orientation), published in 1958, describes three basic needs in every relationship: inclusion (belonging), control (influence) and – in his later work – openness.23 According to Schutz, groups tend to work through these themes in that order: first belonging, then influence, finally openness. FIRO is one of the cornerstones of Sweden's UGL programme and lives on internationally in The Human Element methodology.24

A team holding a retrospective at a whiteboard
Retrospectives and follow-ups are everyday arenas where a group can move from conflict to trust. Photo: Unsplash

Psychological safety and workplace regulation

The phrase "psychological safety" rarely appears in legislation, but managing psychosocial risks is increasingly a legal duty. A few examples:

  • Sweden: the Swedish Work Environment Authority's rules on the organisational and social work environment (formerly AFS 2015:4, part of AFS 2023:2 since 1 January 2025) require employers to prevent unhealthy workloads and victimisation and to set goals for the social work environment.25
  • United Kingdom: the Health and Safety Executive's Management Standards cover demands, control, support, relationships, role and change as the main sources of work-related stress.31
  • International: ISO 45003:2021 gives guidance on managing psychological health and safety at work as part of an occupational health and safety management system.32

For HR, this means that work on trust, communication and relationships in teams is not just "nice to have" – it is part of systematic health and safety work.

Training and tools: UGL, IMGD, FIRO and more

There are many credible organisations working with group development and psychological safety. The right fit depends on whether you want deep personal development, measurement or ongoing everyday practice. Here are some well-established options:

Experiential programme · Sweden

UGL – Developing Groups and Leaders

A classic Swedish residential course (about five days), developed within the Swedish Armed Forces in the 1980s and now owned by the Swedish Defence University. It is delivered by certified facilitators at several providers, such as Folkuniversitetet. Focus on feedback, self-awareness and group dynamics.

fhs.se → · folkuniversitetet.se →
Measuring psychological safety · Global

The Fearless Organization Scan

An assessment developed together with Amy Edmondson and based on her research, with certified practitioners worldwide. In Sweden, certification is offered by the Swedish partner Add Insight.

fearlessorganization.com → · addinsight.se →
Measuring group maturity · Global

IMGD & GDQ

GDQ Associates manages Susan Wheelan's GDQ and trains certified consultants. Swedish consultancies working with group development and psychological safety include Sandahl Partners.

gdq.se → · sandahlpartners.se →
Self-awareness & openness · Global

The Human Element (FIRO Element B)

Will Schutz's further development of FIRO, now stewarded by Henry Thompson. Used in leadership and team development and delivered by certified practitioners in many countries.

thehumanelement.com →
Team behaviours · Global

The Table Group

Patrick Lencioni's organisation, known for The Five Dysfunctions of a Team (2002). Its model puts trust – being able to show vulnerability – at the base of every other team behaviour.33

tablegroup.com →
Game-based everyday practice

Open Windows Game & KYKM

Our own complement: a digital game built on the Johari Window that teams can play again and again in 15–60 minutes, plus the KYKM quick test that maps a team's communication gaps.

openwindowsgame.com → · knowyouknowme.com →

This list is a selection of examples, not a ranking. AB IGLOO Utveckling has no commercial partnership with the external organisations mentioned above.

How to choose the right approach

ApproachBest forTimeStrength
UGLLeaders and individuals seeking deep personal developmentResidential, ~5 daysIntense experience, strong self-insight
GDQ / IMGDLeadership teams wanting to understand their stage of developmentSurvey + workshopResearch-based measurement over time
Fearless Organization ScanOrganisations that want to measure psychological safety specificallySurvey + follow-upDirect link to Edmondson's research
The Human ElementLeaders and teams working on openness and self-esteemMulti-day workshopDeep grounding in FIRO
Open Windows GameWhole teams – including remote and hybrid – wanting regular practice15–60 min per sessionLow threshold, repeatable, measures the "open window" in %

Does team building work at all? A large meta-analysis by Klein and colleagues (2009) found a positive effect – especially for interventions that clarify goals and roles.26 The effect is greatest when the intervention is tied to real work rather than a stand-alone event once a year.

8 ways to build psychological safety every day

Research suggests that safety is built through small, repeated actions – not one-off events. Here are concrete habits you can start next week:

  1. Show that you can be wrong tooLeaders who say "I may have missed something – what do you see?" lower the threshold for everyone else. Edmondson calls this acknowledging your own fallibility.6
  2. Start meetings with a roundLet everyone say something in the first few minutes. People who have spoken once find it easier to speak again – and speaking time becomes more even, which Woolley links to collective intelligence.8
  3. Run the Abilene testHave everyone vote or write down their position before the discussion starts, so nobody is swayed by the first or loudest voice.
  4. Hold a pre-mortemImagine the project has failed a year from now – and ask everyone to write down why. Gary Klein's method makes it legitimate to raise risks.30
  5. Replace "Any questions?" with "What questions do you have?"A small change that assumes questions exist – and makes asking them normal.
  6. Reward the messengerThank people who raise problems, especially uncomfortable ones. It is the single strongest signal of what is safe.
  7. Practise feedback with structureUse the Johari Window as a frame: "Here's what I see in you…" and "Here's something I'd like you to know about me…". Structure makes it safer to begin.
  8. Follow up and measureMeasure safety with a short survey (for example Edmondson's seven items) or a tool such as GDQ, KYKM or Open Windows Game – then do it again a few months later.
A diverse team collaborating with sticky notes
Small, recurring habits build more safety than one big event a year. Photo: Unsplash

From theory to play: the Johari Window in Open Windows Game

Open Windows Game (Empatispelet in Swedish) was created by psychologist Olof Rehn to turn the research above into a concrete, repeatable experience. The game is built directly on the Johari Window:

  • Each player answers a question about themselves – and at the same time guesses how their colleagues answered. This reveals both blind spots and false assumptions, the very mechanism behind the Abilene Paradox.
  • Answers are revealed simultaneously and turns are shared fairly, so everyone is heard – including the quiet ones.
  • After each round, the group's and each individual's "open window" is shown as a percentage, giving you a measure to follow over time.
  • It works on phones, tablets and computers, in the same room or over Teams, Zoom and Google Meet.

The game doesn't replace a UGL course or a professional group development process – but it gives teams a light-hearted way to practise openness and feedback between larger initiatives, week after week.

Colleagues around a table with laptops during a team-building session
A round takes 15–60 minutes and works as an opener for a team meeting, an onboarding session or a workshop.

Frequently asked questions about psychological safety

What does psychological safety mean?

It is a shared belief within a team that it is safe to take interpersonal risks – to ask questions, admit mistakes, offer ideas and disagree – without being embarrassed, rejected or punished. The modern definition comes from Amy Edmondson (1999).1

What is the difference between the Abilene Paradox and groupthink?

In groupthink, the group genuinely believes in the decision. In the Abilene Paradox, most members privately disagree, but nobody says so because everyone assumes the others are in favour. Groupthink is about conviction – Abilene is about silence.

What is UGL, and who delivers it?

UGL (Utveckling av Grupp och Ledare – Developing Groups and Leaders) is an experiential residential course owned by the Swedish Defence University. It is delivered by certified facilitators at various providers, such as Folkuniversitetet. Always check that the facilitators are certified by the Swedish Defence University.

How do you measure psychological safety?

The most common instrument is Amy Edmondson's seven-item scale, which also underpins The Fearless Organization Scan. Wheelan's GDQ measures a group's stage of development, and tools such as KYKM and Open Windows Game reveal the gap between how we see ourselves and how others see us.

Is psychological safety a legal requirement?

Rarely in those words, but many countries require employers to manage psychosocial risks – for example Sweden's AFS 2023:2, the UK HSE Management Standards and the international guideline ISO 45003:2021.253132

Does it work for remote and hybrid teams?

Yes. Google's research found that sitting in the same office was not decisive.5 Remote teams do, however, need more deliberate structures to make sure everyone is heard, because the spontaneous conversations by the coffee machine are missing.

References and further reading

  1. Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350–383. doi.org/10.2307/2666999
  2. Edmondson, A. C. (1996). Learning from mistakes is easier said than done: Group and organizational influences on the detection and correction of human error. The Journal of Applied Behavioral Science, 32(1), 5–28. doi.org/10.1177/0021886396321001
  3. Kahn, W. A. (1990). Psychological conditions of personal engagement and disengagement at work. Academy of Management Journal, 33(4), 692–724. doi.org/10.5465/256287
  4. Frazier, M. L., Fainshmidt, S., Klinger, R. L., Pezeshkan, A., & Vracheva, V. (2017). Psychological safety: A meta-analytic review and extension. Personnel Psychology, 70(1), 113–165. doi.org/10.1111/peps.12183
  5. Google re:Work (2015). Guide: Understand team effectiveness. rework.withgoogle.com
  6. Edmondson, A. C. (2019). The Fearless Organization: Creating Psychological Safety in the Workplace for Learning, Innovation, and Growth. Hoboken, NJ: Wiley.
  7. Duhigg, C. (2016, February 25). What Google learned from its quest to build the perfect team. The New York Times Magazine. nytimes.com
  8. Woolley, A. W., Chabris, C. F., Pentland, A., Hashmi, N., & Malone, T. W. (2010). Evidence for a collective intelligence factor in the performance of human groups. Science, 330(6004), 686–688. doi.org/10.1126/science.1193147
  9. Luft, J., & Ingham, H. (1955). The Johari window, a graphic model of interpersonal awareness. Proceedings of the Western Training Laboratory in Group Development. Los Angeles: UCLA.
  10. Luft, J. (1969). Of Human Interaction. Palo Alto, CA: National Press Books.
  11. Vazire, S. (2010). Who knows what about a person? The self–other knowledge asymmetry (SOKA) model. Journal of Personality and Social Psychology, 98(2), 281–300. doi.org/10.1037/a0017908
  12. Pronin, E., Kruger, J., Savitsky, K., & Ross, L. (2001). You don't know me, but I know you: The illusion of asymmetric insight. Journal of Personality and Social Psychology, 81(4), 639–656. doi.org/10.1037/0022-3514.81.4.639
  13. Gilovich, T., Savitsky, K., & Medvec, V. H. (1998). The illusion of transparency: Biased assessments of others' ability to read one's emotional states. Journal of Personality and Social Psychology, 75(2), 332–346. doi.org/10.1037/0022-3514.75.2.332
  14. Eurich, T. (2018, January 4). What self-awareness really is (and how to cultivate it). Harvard Business Review. hbr.org
  15. Harvey, J. B. (1974). The Abilene paradox: The management of agreement. Organizational Dynamics, 3(1), 63–80. doi.org/10.1016/0090-2616(74)90005-9
  16. Janis, I. L. (1972). Victims of Groupthink: A Psychological Study of Foreign-Policy Decisions and Fiascoes. Boston: Houghton Mifflin.
  17. Ross, L., Greene, D., & House, P. (1977). The "false consensus effect": An egocentric bias in social perception and attribution processes. Journal of Experimental Social Psychology, 13(3), 279–301. doi.org/10.1016/0022-1031(77)90049-X
  18. Morrison, E. W., & Milliken, F. J. (2000). Organizational silence: A barrier to change and development in a pluralistic world. Academy of Management Review, 25(4), 706–725. doi.org/10.5465/amr.2000.3707697
  19. Tuckman, B. W. (1965). Developmental sequence in small groups. Psychological Bulletin, 63(6), 384–399. doi.org/10.1037/h0022100; Tuckman, B. W., & Jensen, M. A. C. (1977). Stages of small-group development revisited. Group & Organization Studies, 2(4), 419–427.
  20. Wheelan, S. A. (2005). Creating Effective Teams: A Guide for Members and Leaders. Thousand Oaks, CA: SAGE.
  21. Wheelan, S. A. (1994). Group Processes: A Developmental Perspective. Boston: Allyn & Bacon.
  22. GDQ Associates. The Group Development Questionnaire (GDQ) and IMGD. gdq.se
  23. Schutz, W. C. (1958). FIRO: A Three-Dimensional Theory of Interpersonal Behavior. New York: Rinehart.
  24. Schutz, W. (1994). The Human Element: Productivity, Self-Esteem, and the Bottom Line. San Francisco: Jossey-Bass. See also thehumanelement.com
  25. Swedish Work Environment Authority (Arbetsmiljöverket). AFS 2023:2 Planning and organisation of work (replacing, among others, AFS 2015:4 Organisational and social work environment, from 1 January 2025). av.se
  26. Klein, C., DiazGranados, D., Salas, E., Le, H., Burke, C. S., Lyons, R., & Goodwin, G. F. (2009). Does team building work? Small Group Research, 40(2), 181–222. doi.org/10.1177/1046496408328821
  27. Collins, N. L., & Miller, L. C. (1994). Self-disclosure and liking: A meta-analytic review. Psychological Bulletin, 116(3), 457–475. doi.org/10.1037/0033-2909.116.3.457
  28. Aron, A., Melinat, E., Aron, E. N., Vallone, R. D., & Bator, R. J. (1997). The experimental generation of interpersonal closeness. Personality and Social Psychology Bulletin, 23(4), 363–377. doi.org/10.1177/0146167297234003
  29. Pentland, A. (2012). The new science of building great teams. Harvard Business Review, 90(4), 60–70. hbr.org
  30. Klein, G. (2007). Performing a project premortem. Harvard Business Review, 85(9), 18–19. hbr.org
  31. Health and Safety Executive (UK). Management Standards for work-related stress. hse.gov.uk
  32. International Organization for Standardization (2021). ISO 45003:2021 Occupational health and safety management – Psychological health and safety at work – Guidelines for managing psychosocial risks. iso.org
  33. Lencioni, P. (2002). The Five Dysfunctions of a Team: A Leadership Fable. San Francisco: Jossey-Bass.

Book a 30-min team demo

Pick a date and time below to schedule a walkthrough with our team.