Psychological Safety: Why 'Feeling Safe' Isn't the Same as 'Being Comfortable'

A review tracing psychological safety from its 1965 origins to its current status as, in Edmondson and Bransby's 2023 assessment, a 'mature theory' with high agreement on definition and measurement, sets out what the construct actually means, what damages it, how it can be recovered once broken, and how it is measured in practice.

What it is — and what it deliberately is not

Edmondson's 1999 definition remains the field's anchor: a shared belief held by team members that the team is safe for interpersonal risk-taking — that speaking up will not lead to embarrassment, rejection or punishment. In short, felt permission for candour. The research is emphatic about what this is not: it is not a 'safe space' shielding people from challenge, not lowered standards or reduced accountability, not unconditional niceness or conflict avoidance, and not a personality trait like introversion. Consequences for poor work still exist. What changes is whether people feel safe to surface problems before they become failures.

A history in five moments

• 1965 — Schein & Bennis argue psychological safety is essential for organisational change, helping people overcome 'learning anxiety'.
• 1990 — William Kahn revives the construct, showing PS enables people to 'employ themselves' at work rather than withdraw.
• 1999 — Amy Edmondson's landmark hospital study finds high-PS teams report more errors — not because they make more, but because they feel safe reporting them.
• 2012 — Google's Project Aristotle studies 180+ teams and finds psychological safety is the single most important factor predicting team effectiveness.
• 2023 — Edmondson & Bransby review 185 empirical studies and declare psychological safety research a mature theory.

What damages it

• Humiliation and ridicule — publicly mocking mistakes creates a powerful chilling effect on everyone watching.
• Blame culture — when errors are punished rather than investigated, people hide problems rather than surface them.
• Supervisor incivility — rudeness cascades down the hierarchy, reducing group-level safety.
• Fear-based leadership and rigid hierarchy — clinicians' psychological safety measurably drops when hierarchy is reinforced.

The silence spiral

Where psychological safety is absent, a self-reinforcing cycle typically emerges: fear of consequences leads an individual to stay silent; others observe this and adopt silence as the norm; problems, errors and improvement ideas go unreported; the organisation loses the information it needs to learn; and unaddressed problems escalate into larger failures. The research cites Boeing and Volkswagen as examples of exactly this pathway playing out at organisational scale.

Recovering it once it's broken

Recovery is described as slow and deliberate, not a single gesture: acknowledging the damage explicitly, genuine accountability without deflection, repeated new experiences where speaking up is actually welcomed, and — critically — structural change where punitive systems or missing voice channels caused the damage in the first place. Cultural work alone will not hold if the structures that broke trust remain in place.

How it's actually measured

Edmondson's original seven-item scale remains the gold standard (e.g. 'if you make a mistake on this team, it is often held against you' — reverse-scored). More recent work, including Plouffe et al.'s (2023) 30-item, five-factor Psychological Safety Inventory, captures a fuller picture across voice safety, failure acceptance, inclusion, respect and fairness, and leader support. Measurement has real limitations, though: social desirability bias means staff in the least safe environments may be least likely to report it honestly, and most surveys capture only a single point in time despite psychological safety being genuinely dynamic.