What Psychological Safety Actually Requires From a Manager

Picture a team meeting. Someone starts to raise a concern about a deadline that everyone privately suspects is unrealistic. Halfway through the sentence they catch the manager’s jaw tighten, and the concern quietly dissolves into “…but I’m sure we’ll be fine.” No one writes that moment down. It won’t appear in any report. But that is the moment psychological safety was actually decided — not in the values deck, not in the engagement survey, but in a half-second facial reaction that the whole team read instantly and filed away.

Psychological safety has become one of the most cited ideas in modern management and one of the most misunderstood. It is talked about as a culture, a climate, a company-wide state of being. In practice it is far more local and far more behavioural than that — and treating it as anything else is why so many well-intentioned efforts to build it change nothing.

What is psychological safety, really?

The term comes from Amy Edmondson, whose 1999 study of 51 work teams defined team psychological safety as “a shared belief held by members of a team that the team is safe for interpersonal risk taking” — the confidence that you won’t be punished or humiliated for speaking up with a question, a concern, an idea, or a mistake. That is the whole of it. It is not a feeling of comfort. It is not the absence of pressure. It is the specific belief that candour is safe.

This is where most of the confusion starts. Edmondson has spent years correcting two misreadings: that psychological safety means being nice, and that it means lowering the bar. It means neither. In her framing, safety sits on one axis and accountability on the other, and only when both are high do you get the “learning zone.” High safety with low accountability isn’t safety at all — it’s a comfort zone where people speak freely and nothing gets done. So the manager’s job is not to make people comfortable. It is to make candour safe while holding standards high. Leaders who suspect that “psychological safety” is code for going soft are half right about the way the term gets abused — and wrong about the thing itself.

Why do psychological-safety initiatives so often fail?

Because organisations try to install it the way they install a policy. A line in the values. An awareness workshop. An item on the annual survey. None of that makes a single person safer to their manager, because safety isn’t a property of the company — it’s a property of the local team, and it is set largely by how one manager behaves. This is why the same organisation, under the same values statement, reliably contains one team where people surface problems early and another, two desks over, where they go silent. The difference isn’t the policy. It’s the person running the room.

Which means the gap is not awareness. Ask almost any manager whether psychological safety matters and they’ll say yes and define it competently. Then watch them shut a person down with a sigh and a “yes, but” the next time an inconvenient point lands at a bad moment. Knowing that safety matters and producing it in real time are different things entirely. Awareness training closes a gap that, for most managers, was never open.

What does the evidence say psychological safety depends on?

Three bodies of evidence point the same way. Edmondson’s original research found that psychological safety predicted whether teams engaged in learning behaviour — asking for help, admitting error, raising problems — and, through that, how they performed; and that leader coaching and behaviour were among its antecedents. Google’s Project Aristotle, which studied 180 of its own teams across 250 attributes, reached a now-famous conclusion: of the five dynamics that separated its most effective teams, psychological safety mattered far more than the rest and underpinned the other four. The best teams weren’t the ones with the most talent; they were the ones where people felt safe to speak up.

And yet McKinsey’s 2021 research found that only 26 percent of leaders were actually creating the conditions for psychological safety on their teams. Read that number carefully. It is not evidence that leaders don’t value safety — the concept has never been more popular. It is evidence of a behaviour gap: the specific things a manager has to do to create safety are mostly not being done, even by people who could explain the idea fluently. The lever is manager behaviour, and most hands aren’t on it.

What does a manager actually have to do to create it?

Edmondson’s own answer is refreshingly concrete, and it is a list of behaviours rather than beliefs. Frame the work as a learning problem, not just an execution problem — something like “this is new for us, we will get things wrong, and our job is to catch them fast” — because if people believe the only acceptable outcome is flawless delivery, they’ll hide anything that isn’t flawless. Acknowledge your own fallibility out loud: “I might be wrong about this — what am I missing?” signals that not-knowing is allowed for everyone. Ask genuine questions and actually listen to the answers, so that curiosity, not interrogation, sets the tone. And then — the one that decides everything — respond productively when someone does take the risk.

That last behaviour is the hinge. Every time a person offers a concern, a challenge, or an admission, the manager’s response either grows safety or erodes it, and the team calibrates in real time. Appreciate the contribution even when you disagree with it. Never punish the messenger for the message. Take visible action on the good suggestions. None of this is an attitude you hold privately; all of it is a thing you do, out loud, in the moment, in front of people who are watching to see whether it’s safe to do it again.

Why is responding well in the moment so hard?

Because the moment that counts is almost never convenient. It’s the challenge raised in front of the team, the mistake that lands squarely on you, the awkward truth delivered at the worst possible time. Under that load, managers default to the reaction that protects them: the tight jaw, the fast rebuttal, the “yes, but” that ends the conversation. It happens in a second, often below conscious control, and the team reads it perfectly.

This is the part that awareness cannot fix. Knowing you should stay open and staying open while you’re irritated, embarrassed, or under pressure are two different capabilities. The first is knowledge you can recall when calm. The second is a performance skill that forms only through repetition — doing the hard response, under something like real conditions, enough times that it becomes the thing that shows up automatically. It is the same reason a manager can recite a feedback framework perfectly and still freeze the instant the other person reacts. The knowledge was never the bottleneck.

What should an L&D or People leader do differently?

Stop trying to install psychological safety through awareness, and start building the behaviours that create it through practice. Concretely, that means three moves. Name the specific micro-behaviours — framing work as learning, admitting your own fallibility, asking real questions, responding well to bad news — rather than leaving “be more open” as an aspiration. Give managers repeated, realistic reps at the hardest one, responding well in the seconds after someone takes an interpersonal risk, somewhere getting it wrong costs nothing. And measure the behaviour where it actually happens — did the manager invite the concern, and what did they do with it — rather than inferring safety from an annual survey that reports the temperature months after the moments that set it.

This is the logic HumanVantage is built on. Psychological safety is a set of manager behaviours performed under pressure, and behaviours change through rehearsal, not exposure. Managers practise the specific moments — a challenge raised in a meeting, a mistake confessed, a hard piece of feedback handed back — against realistic reactions and with unlimited retries, so that the productive response becomes the one that surfaces when it counts, and the change is visible at the point of practice rather than guessed at from a survey.

If you want more psychological safety on your teams, the least effective thing you can do is add it to the values page. Pick the handful of manager behaviours that create it, and make them practisable. The manager who has rehearsed staying open to a challenge twenty times will stay open on the twenty-first — the one that’s real, inconvenient, and in front of everyone.