Knowledge base Updated 2 November 2026 1 views

Psychological Safety

A shared belief within a team that it is safe to take interpersonal risks — to ask a question, admit a mistake, disagree or raise a concern — without being humiliated or penalised for it.

Definition

Psychological safety is a property of a team climate, not of an individual and not of an organisation as a whole. It varies substantially between teams inside the same company, which is one of the better-replicated findings about it.

It is not the same as comfort, trust, or low standards. Edmondson's own framing pairs it with accountability: high safety with low standards produces a comfort zone, high standards with low safety produces anxiety, and learning requires both.

History

The term appears in Schein and Bennis's work on organisational change in 1965. Amy Edmondson gave it its modern form in a 1999 study of hospital teams.

That study produced a result that looked backwards: better-performing units reported more medication errors. The explanation was that they reported more because it was safe to, not that they made more. That inversion — a safety measure that reads as a failure measure — is the finding the whole construct rests on and is still the clearest illustration of it.

Google's Project Aristotle identified it as the strongest predictor among the factors it examined, which is why it is now widely known outside research.

Theory

The mechanism is the reduction of interpersonal risk. Speaking up about a problem carries a predictable cost — looking ignorant, incompetent, negative or disruptive — and silence is the individually rational response when that cost is high. The cost is paid by the organisation, in problems that are known and unreported.

Safety therefore matters most where work is uncertain and interdependent, and matters least for routine independent tasks. This is a moderator with reasonable support and it is usually dropped from popular accounts, which present safety as universally beneficial.

Leader behaviour is the strongest lever: inclusiveness, acknowledging fallibility, and inviting input.

Research

Meta-analytic work links psychological safety to learning behaviour, information sharing, creativity and performance, with the association to learning behaviour stronger and more consistent than the association to performance.

In healthcare it predicts error reporting and speaking up about patient safety, and in aviation it is the psychological substance behind crew resource management.

The methodological weaknesses are real: much of the literature is cross-sectional and self-report, and psychological safety and the outcomes are frequently measured from the same respondents at the same time, which inflates associations. Longitudinal and intervention evidence is thinner than the volume of publication suggests.

Applications

Team design and leadership development, clinical governance and incident reporting, safety-critical operations, and post-incident review.

The things that change it are structural and specific: how a leader responds the first time someone raises bad news, whether reporting a near miss has consequences, whether meetings have a mechanism for dissent, and whether the person who raised a problem hears what happened next.

Training alone does not produce it. A leader who runs a psychological safety workshop and then punishes the next piece of bad news has taught the opposite lesson more effectively than the workshop taught anything.

Criticism

Cross-sectional and common-method evidence dominates, so causal claims outrun the design.
Popularisation has flattened it. It is frequently presented as being nice, which drops the accountability half and makes it unfalsifiable.
Level-of-analysis confusion. It is a team construct routinely measured and discussed as an individual one.
Overlap with trust, voice climate and leader-member exchange is substantial and discriminant validity is not well established.
It can be used to individualise structural problems: telling people they are safe to speak up while leaving intact the hierarchy that makes it unsafe.

References

Edmondson, A. C. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350-383.
Edmondson, A. C., & Lei, Z. (2014). Psychological safety: The history, renaissance, and future of an interpersonal construct. Annual Review of Organizational Psychology, 1, 23-43.
Frazier, M. L., et al. (2017). Psychological safety: A meta-analytic review and extension. Personnel Psychology, 70(1), 113-165.
Newman, A., Donohue, R., & Eva, N. (2017). Psychological safety: A systematic review of the literature. Human Resource Management Review, 27(3), 521-535.

Written for PsychologyOrbit and reviewed before publishing. It is a summary for professionals and students, not clinical guidance, and it is not a substitute for the primary sources listed above.