One of the teams I worked most closely with over several years had a leader who genuinely cared about the people on her team and worked hard to ensure they felt supported. She checked in on team members regularly, responded quickly to personal difficulties, protected the team from organizational pressures that she could absorb, and ran meetings with a tone that was consistently warm and encouraging. By almost any observable measure, it was a pleasant team to work on. It was also, by any rigorous measure, an average-performing team — one that consistently delivered what was expected and consistently failed to identify problems before they became serious, adapt to changing conditions, or challenge its own practices in any meaningful way.
When I spent time with the team trying to understand the gap, what I found was specific. The team members were comfortable — genuinely comfortable with each other and with their leader. But they had learned, without anyone saying so explicitly, that the kind of things that would have made them less comfortable — flagging a serious problem with an initiative the leader was excited about, raising concerns about another team member's performance, questioning a direction that the team had already collectively committed to — would not be received well. Not badly, exactly, but in ways that created awkwardness, required careful management, and generally made the person who raised the uncomfortable thing feel that they had caused a problem. Over time, that implicit signal had shaped what people brought to team conversations. The team was psychologically comfortable. It was not psychologically safe.
The distinction is the core of what psychological safety actually means and why it matters. Psychological safety is not about whether people enjoy working with each other, feel supported by their leader, or find the work environment pleasant. It's the specific condition where people believe that raising difficult information — including information that challenges the consensus, threatens the team's self-image, or creates awkward conversations — will not result in interpersonal punishment. That belief is different from comfort, and the two are sometimes in tension: teams that are very comfortable can develop strong implicit norms against the disruption of that comfort, which directly undermines psychological safety.
What psychological safety actually enables
The organizational value of psychological safety is not that it makes teams feel good about working together, though it typically does. The value is informational: psychologically safe teams surface problems earlier, surface them more completely, and produce better analysis of complex situations because the people doing the analysis aren't filtering what they surface based on how it will be received.
Every team has members who have observed problems, held minority views, or noticed information that complicates the prevailing picture. In teams with low psychological safety, that information stays with the individuals who hold it — either because they've learned that surfacing it is socially costly, or because they've rationalized it away by accepting the consensus view. In teams with high psychological safety, that information enters the group's deliberation and improves its quality. The functional difference is significant: teams with low psychological safety are operating on a systematically incomplete picture of their situation, because the filtering that happens at the individual level before information reaches the group is removing precisely the information that would be most useful — the qualifying data, the contrary evidence, the early warning signals.
The second functional value is in learning. Teams that can honestly examine what went wrong — and where each person's specific actions contributed to outcomes that didn't go as planned — extract learning from experience that teams without that honesty cannot. The retrospective that produces genuine insight requires the ability to say "what I did in this situation made it worse, here's how" without the social cost of that admission being prohibitive. The learning discipline that high-performing teams build is impossible without a sufficient level of psychological safety to make honest retrospection socially viable.
The accountability pairing
Psychological safety without high accountability produces the comfort zone I described in the opening: teams that are pleasant, low-conflict, and plateaued. The performance research consistently finds that the combination that produces both excellent performance and high wellbeing is high psychological safety paired with high accountability — what Amy Edmondson, who originally formalized the concept, describes as a "learning zone."
This pairing is difficult to maintain because the two conditions can feel like they're in tension. High accountability — clear expectations, visible tracking of commitments, genuine consequences for consistent underperformance — can feel threatening, and in environments with low psychological safety, it is: when people aren't confident they can raise problems without punishment, high accountability means people will be punished for problems they should have felt safe surfacing earlier. The sequence that resolves this tension is specific: you build psychological safety first, or simultaneously with accountability, never after. High accountability in an environment with low psychological safety produces the "anxiety zone" — people working hard, taking few risks, and hiding problems until they're no longer hideable.
The practical implication for leaders is that raising accountability without attending to the safety conditions is a predictable path to a specific kind of organizational dysfunction: people who look like they're performing by every visible metric while actually operating in a mode that will produce significant problems when something genuinely difficult happens. The leader who wants high accountability on a team needs to attend with equal energy to whether the safety conditions for genuine accountability are present.
How leaders build it — and how they accidentally destroy it
Psychological safety is built primarily through the leader's repeated, consistent modeling of the behavior it requires. The leader who receives difficult information without visible defensiveness — who treats a problem surfaced as useful information rather than as an accusation, who acknowledges their own errors openly and without excessive self-flagellation, who visibly changes their thinking when challenged well — is demonstrating through behavior that the safety conditions are real. This modeling works because team members are watching the leader's responses to determine whether the implicit norms match the stated ones.
The leader doesn't need to do this perfectly; they need to do it consistently enough that the signal is clear. A leader who usually receives difficult information well but occasionally responds in ways that punish the messenger creates the worst of both worlds: team members have no reliable sense of when it's safe to surface difficult things, which typically produces a strategy of not surfacing them. The consistency of the leader's modeling is more important than its perfection.
The most common inadvertent destroyers of psychological safety are not hostile responses but ambiguous ones. The leader who doesn't visibly engage with the difficult thing that was raised — who moves on quickly, who doesn't acknowledge it in subsequent conversations, who uses the messenger's concern against them in some subtle way later — teaches team members that surfacing difficult information produces a worse experience than not surfacing it, even if no one is explicitly punished. The distinction between accountability and blame is directly relevant here: environments where raising problems leads to the problem-raiser being associated with the problem they raised are environments where people learn not to raise problems.
The diagnostic practice
The most reliable diagnostic for psychological safety on a team is not a survey — it's direct observation of what happens when difficult information is surfaced. Does the team engage with it as information, or manage it as a social problem? Do people raise concerns during the discussion, or after the meeting when the leader isn't there? When something went wrong, does the post-mortem produce genuine analysis, or careful attribution management? These observable behaviors reveal the actual safety level more reliably than any stated belief about the team culture.
A second diagnostic is to notice what's not being said. The experienced leader can often feel the absence of difficult information in team conversations — the meetings that go too smoothly, the decisions that get endorsed without the objections that the leader knows some team members hold. When the meeting experience and the hallway-conversation experience diverge significantly — when the real discussions happen outside the formal settings — psychological safety is almost certainly insufficient for the conversations the team actually needs to have. The gap between what is said in the room and what is said after the meeting is one of the most reliable indicators of safety deficit in any team.
Closing that gap is the work of psychological safety development. It doesn't happen through team retreats or stated commitments to candor; it happens through the repeated experience of difficult information being surfaced and receiving a response that confirms the surface was worth making. The skill of delivering difficult messages is the complement to the leader's skill of receiving them: both need to be present for the informational value of psychological safety to be realized.