Psychological Safety: Definition, Research, and Practical Guide for Teams

A product development team at a major automotive manufacturer discovers a design flaw during the testing phase. The engineer who finds it stays silent — because he fears being perceived as the person who delays the project. The flaw is discovered six months later. Cost: 4.2 million euros and a missed market window.

This is not an isolated case. It is a system characteristic. And it has a name: absent psychological safety.

The term sounds soft. The research behind it is hard. Google studied 180 teams over two years and found: psychological safety is by far the strongest predictor of team performance — stronger than team composition, resources, or individual expertise.¹ Amy Edmondson at Harvard Business School has empirically grounded the concept over 25 years.² And yet most organizations in the DACH region either ignore it or dismiss it as a “feel-good topic.”

This article explains what psychological safety actually is (and what it is not), what research supports it, how you can measure and build it — and why teams without psychological safety cannot innovate. The connection to innovation culture and business transformation is no coincidence: psychological safety is not a nice addition. It is the prerequisite.

What Is Psychological Safety?

Psychological safety is the shared belief within a team that interpersonal risks — such as asking questions, admitting mistakes, voicing ideas, or raising concerns — will not be punished.

Amy Edmondson defined the construct in 1999 in her study of hospital teams: “Psychological safety is a shared belief held by members of a team that the team is safe for interpersonal risk taking.”² Her original observation was paradoxical: the teams with the best leaders reported more errors, not fewer. Not because they made more mistakes — but because they worked in a climate where errors could be openly addressed.

The key word is “shared.” Psychological safety is not an individual trait (that would be self-confidence). It is a collective perception at the team level: “In this team, I can show up without fear of consequences.” It does not arise from a single measure but from repeated experience — every time someone makes a vulnerable statement and receives no negative reaction, the collective sense of safety increases.

What Psychological Safety Is Not

Three common misconceptions are dangerous because they either dilute or discredit the concept:

1. It is not “being nice to each other.” Psychological safety does not mean the absence of conflict. On the contrary: teams with high psychological safety have more open disagreements — because they dare to voice dissent. It is not about avoiding uncomfortable truths but about being able to speak them.

2. It is not the absence of performance standards. Edmondson herself emphasizes: psychological safety without accountability creates a “comfort zone.” Teams need both — the safety to take risks and clear expectations for results. The combination of high psychological safety and high performance standards creates what Edmondson calls the “learning zone.”³

3. It is not trust. Trust is a bilateral relationship between two people (“I trust you”). Psychological safety is a property of the team as a system (“In this team, it is safe to be vulnerable”). A team can consist of people who trust each other and still lack psychological safety — if group norms sanction the open addressing of problems.

The Research Base

Edmondson’s Hospital Study (1999)

Edmondson studied medication errors in hospital units. The hypothesis: better teams make fewer errors. The result was surprising: the best teams reported the most errors. The explanation: in psychologically safe teams, errors are reported and addressed. In unsafe teams, they are hidden — with potentially fatal consequences for patients.²

This study founded the entire research field. Over 200 empirical studies in the following 25 years confirmed the relationship between psychological safety and team performance, innovation behavior, and learning capacity.⁴

Google’s Project Aristotle (2015)

Starting in 2012, Google studied 180 teams under the codename “Project Aristotle” to determine what distinguishes high-performing teams from average ones. Researchers analyzed over 200 interviews and 250 team attributes — from personality traits to educational backgrounds to work habits.¹

The result was unambiguous: psychological safety was the strongest differentiator. Teams with high psychological safety exceeded revenue targets 17 percent more often. Teams with low psychological safety fell 19 percent short of revenue targets. The composition of the team — who was on the team — was less decisive than the team dynamics — how the team worked together.

The four additional factors Google identified: dependability (can team members rely on each other?), structure and clarity (are roles and goals clear?), meaning (does the work have personal relevance?), and impact (does the team believe its work matters?). Psychological safety was the foundation for all four.

Meta-Analyses and Impact Evidence

Frazier et al. (2017) analyzed 136 studies with over 22,000 participants in a comprehensive meta-analysis. Results: psychological safety correlates significantly positively with job satisfaction, commitment, learning behavior, and team performance. It correlates significantly negatively with apathy and interpersonal conflicts.⁴

Newman, Donohue, and Eva (2017) confirmed in a separate meta-analysis: the relationship between psychological safety and performance is robust across different industries, cultures, and hierarchy levels.⁵ This is not a culture-specific phenomenon — it is a universal feature of effective collaboration.

The Four Stages of Psychological Safety

Timothy R. Clark expanded Edmondson’s concept in 2020 into a developmental model with four sequential stages.⁶ The sequence is not arbitrary — each stage builds on the previous one. Attempting to establish Stage 3 (Contributor Safety) without securing Stage 1 (Inclusion Safety) is building on sand.

Stage 1: Inclusion Safety

Definition: Team members feel accepted as belonging — regardless of background, opinion, or status. Basic needs for belonging and connection are met.

Guiding question: “Can I be here?”

Indicators: New team members are actively integrated. Different perspectives are perceived as enrichment. Nobody is systematically excluded or ignored. Informal networks are permeable.

Typical violation: The new colleague is systematically ignored at lunch. The introverted engineer is overlooked in meetings. The colleague with a migration background is not invited to informal after-work gatherings.

Stage 2: Learner Safety

Definition: Team members can ask questions, admit mistakes, and experiment without being punished or shamed. The learning process is recognized as normal and valuable.

Guiding question: “Can I learn?”

Indicators: Questions are taken seriously, not dismissed. Mistakes lead to analysis, not blame. Experimentation is rewarded. “I don’t know” is an accepted answer.

Typical violation: “You should know that by now.” “We have discussed this question three times already.” Eye-rolling at clarifying questions. Mistakes are linked to names: “The Miller mistake.”

Stage 3: Contributor Safety

Definition: Team members can independently contribute, bring ideas, and deploy their competencies — with the certainty that their contribution will be heard and taken seriously.

Guiding question: “Can I contribute?”

Indicators: Ideas are evaluated by quality, not origin. Even unfinished thoughts can be shared. Contributions from all hierarchy levels are weighted equally. There are spaces for creative impulses.

Typical violation: Only contributions from seniors are discussed. Ideas are ignored and presented by someone else as their own weeks later. “That does not work here” as the standard response to new proposals.

Stage 4: Challenger Safety

Definition: Team members can question the status quo, speak uncomfortable truths, and challenge existing decisions — without fear of retaliation.

Guiding question: “Can I disagree?”

Indicators: Criticism of decisions is valued as constructive input. “Devil’s advocate” roles are institutionalized. Leaders respond to disagreement with curiosity rather than defensiveness. Whistleblowing is protected, not sanctioned.

Typical violation: Whoever contradicts the boss gets the unpopular projects. Criticism of strategy is treated as disloyalty. “We have decided this, now we execute.”

Why sequence matters: A team that aims for Stage 4 but has not secured Stage 1 creates a facade of openness. The leader demands “critical thinking” — but half the team does not even dare to ask a question. The result: the self-confident dominate, the others fall silent. The diversity of perspectives — the actual value of psychological safety — is lost.

How to Measure Psychological Safety

Edmondson’s 7-Item Scale

Edmondson developed a validated scale with seven statements rated on a Likert scale (1 = “strongly disagree” to 7 = “strongly agree”).² Three items are positively framed, four are negatively framed (reverse scored):

1. If you make a mistake on this team, it is held against you. (reversed)
2. Members of this team are able to bring up problems and tough issues.
3. People on this team sometimes reject others for being different. (reversed)
4. It is safe to take a risk on this team.
5. It is difficult to ask other members of this team for help. (reversed)
6. No one on this team would deliberately undermine my work.
7. My unique skills and talents are valued and utilized in this team.

Scoring: The average across all items (after reverse coding negative items) produces a team score. Scores above 5.0 are considered solid. Below 3.5 indicates urgent need for action. The diagnostic value lies less in the absolute score than in variance between items and trajectory over time.

Important: The scale is evaluated at team level, not individually. Psychological safety is a group phenomenon. When large differences in ratings exist within a team, this indicates subgroups with different experiences — a signal that Inclusion Safety (Stage 1) is not universally present.

Supplementary Diagnostics

The 7-item scale is a starting point. For deeper diagnosis, supplementary methods can be deployed:

Team retrospectives: Structured reflection rounds that explicitly ask about moments when team members felt unsafe. The quality of answers is itself an indicator: if nobody names such a moment, it can either mean psychological safety is high — or that it is so low that nobody dares say otherwise.

Behavioral audit: Observation of meetings over two to four weeks. Who speaks how much? Are questions asked? How are mistakes handled? How is disagreement treated? The distribution of speaking time is a particularly revealing indicator: in teams with high psychological safety, speaking time is more evenly distributed than in unsafe teams.¹

How to Build Psychological Safety: Leader Behaviors

Psychological safety does not emerge from workshops or posters on the wall. It emerges from repeated leader behavior — especially in moments when it is uncomfortable. Edmondson identifies three core behaviors.³

1. Model Fallibility

What it means: The leader openly admits that they do not know everything and make mistakes. Not as performative humility but as authentic practice.

Specifically: “I made a mistake in budget planning — the licensing costs for Q3 were twice as high as estimated. Here is how it happened, and here is what I learned.” Not: “We made a small mistake there” (passive formulation that diffuses responsibility).

Why it works: When the person with the highest power position shows vulnerability, the perceived danger drops for everyone else. It sets a norm: in this team, it is safe to not be perfect. A study by Nembhard and Edmondson (2006) in surgical teams shows: the willingness of nurses to voice concerns increases significantly when surgeons explicitly name their own fallibility.⁷

2. Frame Work as a Learning Problem

What it means: The leader explicitly communicates that the task contains uncertainty and that mistakes are expected parts of the learning process.

Specifically: “We are entering new territory with this project. I do not expect the first draft to be perfect. I expect us to learn quickly.” Not: “This needs to be right the first time” (creates fear of mistakes and prevents experimentation).

Why it works: When work is framed as a performance problem (“Deliver results”), mistakes become failures. When it is framed as a learning problem (“We learn and improve”), mistakes become data points. The framing determines whether people take risks or avoid them.

3. Actively Invite Participation

What it means: The leader specifically asks for opinions, concerns, and ideas — especially from people who do not volunteer on their own.

Specifically: “Sarah, you have observed the customer process most closely — what do you see that we are missing?” Not: “Does anyone else have something?” (open questions to the group are answered by the most self-confident and interpreted as rhetorical by the uncertain).

Why it works: Active invitation signals: your perspective has value. It breaks the spiral of silence in which those who know the most (because they work closest to operational reality) say the least.

The Feedback Window

A particularly critical moment: how the leader responds to the first vulnerable statement. When someone admits a mistake or asks a critical question for the first time, the entire team watches the reaction.

If the reaction is positive (gratitude, curiosity, constructive discussion): the norm is set. Others will follow.

If the reaction is negative (impatience, eye-rolling, dismissal): the norm is also set. But in the other direction. And correction takes much longer than the violation.

This window does not open often. It is asymmetric: a single negative reaction outweighs ten positive ones. Edmondson calls this the “amplification effect” — negative experiences are weighted more heavily than positive ones regarding psychological safety.³

The Connection to Innovation Culture

Psychological safety is not a supplement to innovation culture — it is its prerequisite. The logic is direct:

Innovation requires experimentation. Experimentation produces mistakes. If mistakes are punished, people stop experimenting. No experiments, no innovation.

Innovation requires dissent. Breakthroughs happen when someone says: “What if we do this completely differently?” If dissent is considered disloyalty, nobody says that sentence. No uncomfortable questions, no breakthroughs.

Innovation requires knowledge transfer. When an engineer makes a mistake and learns from it, the team only benefits if the knowledge is shared. In teams without psychological safety, mistakes are hidden instead of shared. The team repeats the same mistake — just different people make it at different times.

Practice Example: R&D Department at an Automotive Corporation

An automotive manufacturer discovered that its R&D teams, despite high budgets and qualified staff, produced fewer patent filings than the industry average. Internal analysis revealed: it was not ideas that were missing. It was the willingness to share unfinished ideas.

The cause: a culture of “perfection presentation” — every idea had to be fully developed before it could be presented. This produced two effects: first, ideas were only shared when they were “safe” — meaning incremental rather than radical. Second, ideas that failed early were never communicated — other teams repeated the same dead ends.

The intervention: introduction of “Fail Forward Fridays” — a two-hour weekly format where teams exclusively presented failed experiments and unfinished ideas. The first four weeks were slow (hardly any voluntary participation). Leaders went first and presented their biggest misjudgments of the quarter. After three months, the format was fully booked. After one year, patent filings rose by 23 percent — not because more ideas emerged, but because existing ideas were shared earlier and iterated faster.

Psychological Safety and Transformation

Why transformations fail often has less to do with strategy than with the climate in which it is implemented. The change management models describe the tools. Psychological safety is the operating temperature that determines whether the tools work.

During transformation: Employees must be able to communicate uncertainty. They must be able to say: “I do not understand why we are doing this.” They must be able to report implementation mistakes without being labeled as transformation blockers. Without psychological safety, information does not flow upward — and leadership steers blind.

After transformation: The new structures must be lived. If old behavioral patterns — silence, adaptation, waiting — persist, the new structures are decoration. Psychological safety ensures that employees actually use the new degrees of freedom instead of carrying old caution into the new structure.

Common Mistakes When Building Psychological Safety

Mistake 1: Treating Psychological Safety as a One-Time Initiative

A workshop on psychological safety can create awareness. It does not create psychological safety. That emerges through hundreds of small interactions over months and years. Running a workshop and then checking the box means spending money but changing nothing.

Mistake 2: “Open Door” Policy as Substitute

“My door is always open” is one of the most frequently cited — and least effective — leadership statements. The problem is not the door. The problem is power distance. The leader must actively go to employees — not wait for them to come. In hierarchical organizational cultures, as common in many DACH companies, the threshold for actually using an open door is higher than leaders assume.

Mistake 3: Confusing Psychological Safety with “Harmony”

Teams with high psychological safety argue more, not less. But they argue productively — about ideas, not about people. Leaders who equate conflict avoidance with psychological safety create apparent calm beneath which problems accumulate.

Mistake 4: Expecting Immediate Results

Trust builds slowly and destroys quickly. Research shows: it takes months of consistent leader behavior before a team’s perceived psychological safety changes measurably. A single inconsistent reaction — an eye-roll in a meeting, a sarcastic remark after a mistake — can undo weeks of positive work.

Mistake 5: Focusing Only on the Leader

The leader has the strongest influence but is not the only source. Team members shape psychological safety among themselves through their reactions to mistakes, questions, and ideas. Peer-to-peer dynamics — a dominant team member who talks down to others, or a clique that excludes outsiders — can sabotage the leader’s efforts.

Psychological Safety vs. Related Concepts

Concept	Focus	Level	Relationship to Psychological Safety
Trust	Bilateral relationship	Dyad	Necessary but not sufficient. Trust is individual; psychological safety is a team property.
Error Culture	Handling mistakes	Organization	Psychological safety is the prerequisite for a functional error culture. Without it, error culture remains lip service.
Engagement	Emotional commitment	Individual	Psychological safety increases engagement but is not identical to it. You can be engaged and still feel unsafe.
Resilience	Handling setbacks	Individual/Team	Psychologically safe teams recover faster from setbacks because they discuss errors openly and correct faster.
Servant Leadership	Serving leadership posture	Leader	Servant leadership creates the conditions for psychological safety. The leader places the team at the center.

Cultural Dimension: Psychological Safety in the DACH Region

DACH corporate culture has specific characteristics that both hinder and facilitate building psychological safety.

Hindering:

• Hierarchy orientation. Hofstede’s Power Distance Index shows: Germany, Austria, and Switzerland have moderate power distance, but implicit hierarchies in corporations are often stronger than formal ones. The department head who communicates “at eye level” but whose informal power everyone senses creates a double message.
• Perfection culture. “Zero defect tolerance” as a quality standard is deeply embedded in German industrial culture — and stands in direct contradiction to Learner Safety. The paradox: the quality that made Germany famous can become an innovation brake when it suppresses willingness to experiment.
• Engineering culture. In technology-driven industries, competence is demonstrated through facts. “I don’t know” is interpreted as a competence deficit, not an honest statement.

Facilitating:

• Co-determination. Institutionalized employee co-determination (works councils, supervisory boards) creates structural channels for dissent that other cultures lack.
• Objectivity. The German discussion culture — argument beats hierarchy — provides a cultural anchor for Challenger Safety when taken seriously.
• Long-term orientation. The willingness to invest long-term in employees (apprenticeship system, low turnover rates) creates stable teams in which psychological safety can develop over time.

Frequently Asked Questions

Is psychological safety measurable?

Yes. Edmondson’s 7-item scale has been validated since 1999 and deployed in over 200 studies. Measurement occurs at team level through anonymous surveys. Additionally, behavioral audits (speaking time distribution, error reporting frequency) can serve as objective indicators. A single workshop is insufficient as a measurement instrument — psychological safety should be assessed regularly to capture changes over time.

Can introverted team members experience psychological safety?

Yes — and they benefit particularly from it. Introverted team members often have valuable perspectives that go unheard in unsafe teams because extroverts dominate the space. Psychological safety does not mean everyone must speak equally. It means everyone has the opportunity to be heard — including through written formats, one-on-one conversations, or asynchronous channels.

What to do when the leader is the problem?

This is the most difficult case — and the most common. When the leader actively destroys psychological safety through their behavior, no team interventions help. Options: 360-degree feedback with specific behavioral examples, executive coaching focused on impact awareness, or — in persistent cases — a leadership change. Meta-analyses show: the leader’s influence on psychological safety is so dominant that no team measure can compensate for it.

How long does it take to build psychological safety?

Months, not weeks. For teams that previously worked in a fear culture, often a year or longer. The asymmetry is decisive: building trust is slow and linear. Destroying trust is fast and exponential. One sarcastic remark from the leader after a mistake can undo three months of positive work. Consistency matters more than intensity.

Does psychological safety work in remote teams?

Yes, but it requires more deliberate design. In-person, trust-building micro-interactions happen incidentally (coffee kitchen, shared lunch). Remote, this must be explicitly created: check-ins at the start of meetings, actively inviting quiet participants, deliberate use of camera and chat. Research shows: remote teams can achieve equally high psychological safety as in-person teams — if the leader steers this consciously.

Methodology and Sources

This article is based on 12 academic sources on psychological safety, team research, and organizational psychology. The primary source is Amy Edmondson’s research program (1999—2024), supplemented by Google’s Project Aristotle (2015), two meta-analyses (Frazier et al. 2017, Newman et al. 2017), and Timothy Clark’s 4-stage model (2020).

Limitations: Most studies originate from Western contexts. Cultural transferability to strongly hierarchical organizational cultures in the DACH region is plausible but not fully empirically validated. Practice examples are anonymized and constructed based on typical industry patterns.

Disclosure: SI Labs accompanies organizations through transformation processes in which psychological safety plays a central role. We have endeavored to base recommendations on published research and to transparently acknowledge the limitations of the concept.

Sources

Footnotes

1. Duhigg, Charles. “What Google Learned from Its Quest to Build the Perfect Team.” The New York Times Magazine, February 25, 2016. Based on Google’s internal research “Project Aristotle” (2012—2015). ↩ ↩² ↩³

2. Edmondson, Amy C. “Psychological Safety and Learning Behavior in Work Teams.” Administrative Science Quarterly 44, no. 2 (1999): 350—383. ↩ ↩² ↩³ ↩⁴

3. Edmondson, Amy C. The Fearless Organization: Creating Psychological Safety in the Workplace for Learning, Innovation, and Growth. Hoboken: Wiley, 2019. ↩ ↩² ↩³

4. Frazier, M. Lance, Stav Fainshmidt, Ryan L. Klinger, Amir Pezeshkan, and Veselina Vracheva. “Psychological Safety: A Meta-Analytic Review and Extension.” Personnel Psychology 70, no. 1 (2017): 113—165. ↩ ↩²

5. Newman, Alexander, Ross Donohue, and Nathan Eva. “Psychological Safety: A Systematic Review of the Literature.” Human Resource Management Review 27, no. 3 (2017): 521—535. ↩

6. Clark, Timothy R. The 4 Stages of Psychological Safety: Defining the Path to Inclusion and Innovation. Oakland: Berrett-Koehler, 2020. ↩

7. Nembhard, Ingrid M. and Amy C. Edmondson. “Making It Safe: The Effects of Leader Inclusiveness and Professional Status on Psychological Safety and Improvement Efforts in Health Care Teams.” Journal of Organizational Behavior 27, no. 7 (2006): 941—966. ↩