Psychological safety at work: what it is, what it is not, and how to build it

What the research actually says

Edmondson's original hospital studies found that better-performing nursing teams reported more errors, not fewer. They were not making more mistakes; they were surfacing them. That distinction sits at the heart of psychological safety: it determines whether an organisation learns from what goes wrong or discovers it too late.

Edmondson's 1999 field study of 51 work teams in one manufacturing company found that psychological safety was associated with learning behaviour, which was associated with team performance. Google's internal Project Aristotle studied 180 teams and identified psychological safety among five dynamics of effective teams. These findings do not establish a universal strongest predictor across workplaces.

Four things psychological safety is not

Most resistance to the concept comes from a misunderstanding of what it requires.

• It is not niceness. Safe teams disagree openly and often more directly.
• It is not lowered standards. Safety plus high standards is the learning zone; safety without standards is a comfort zone.
• It is not a personality trait. It is a property of a team, shaped mainly by the leader.
• It is not permanent. It can be lost in a single meeting and takes far longer to rebuild.

How to measure it credibly

Use a short validated scale rather than a single question. Edmondson's seven-item measure covers whether mistakes are held against people, whether difficult issues can be raised, whether difference is accepted, and whether asking for help is safe.

Measure at team level and never report results for groups small enough to identify individuals. Trend matters more than absolute score: a team moving from 3.2 to 3.7 over two quarters tells you the leader behaviours are landing.

The leader behaviours that move the score

Psychological safety is built in small, repeated moments rather than in workshops. Three behaviours carry most of the effect.

• Frame the work as a learning problem: say explicitly that the situation is uncertain and that input is needed
• Model fallibility: name your own mistakes and what you learned, briefly and without theatre
• Respond productively: thank the person who raised the problem before discussing the problem itself
• Close the loop: tell people what happened as a result of what they raised

Where psychological safety meets psychosocial risk

In Germany, the Gefährdungsbeurteilung psychischer Belastung requires employers to assess psychosocial risk. Psychological safety is not a legal requirement in itself, but it determines whether that assessment produces honest data. Teams that do not feel safe will complete the survey defensively, and the organisation will act on a distorted picture.

A 90-day starting plan

Baseline the seven-item measure at team level. Run a 90-minute leader session on the three behaviours above. Ask every team leader to introduce one recurring practice, such as a standing agenda item for what went wrong this fortnight. Re-measure after one quarter and discuss the results with the teams themselves rather than only with senior management.

Frequently asked questions

How is psychological safety different from trust?

Trust is an expectation about another person. Psychological safety is a shared belief held by a group about the climate of the team. A person can trust a colleague individually while still believing it is unsafe to raise a concern in a team meeting.

Can psychological safety lower performance standards?

No. In Edmondson's framing, safety combined with high standards produces the learning zone. Safety without standards produces complacency, and standards without safety produce anxiety and hidden problems.

How often should psychological safety be measured?

Twice a year at team level is usually sufficient, with a short pulse in between if a team is going through significant change. More frequent measurement without visible action reduces credibility.

What if a team scores low?

Treat a low score as information rather than as a verdict on the leader. Discuss it with the team, agree one or two concrete changes, and re-measure after a quarter.