Psychological safety is the shared belief that a team is safe for interpersonal risk-taking: that you can raise a concern, admit a mistake, or challenge a plan without being humiliated, sidelined, or punished. The term entered mainstream management vocabulary through Amy Edmondson’s research on hospital teams, and it has since become a staple of leadership retreats, pulse surveys, and laminated values posters. But on the ground, psychological safety is not a feeling that can be installed through a half-day workshop. It is a property of the material and social structure of work — the meeting cadence, the performance review rubric, the physical layout of a room, the unspoken rule about who speaks first. This article examines why structural change, not training, is the real lever for psychological safety, and what that looks like in the ordinary artifacts of organizational life.
For readers of this blog, the question is not whether psychological safety matters. The more interesting question is what it is made of. In workplace ethnography, we treat psychological safety as an observable pattern of interaction, not an internal state. We look for the traces it leaves: who interrupts whom, which questions get asked in public versus private, whose name appears in the meeting notes, which doors stay open. Those traces reveal the actual rules of the game, and they rarely change because someone attended a session on active listening.
The Workshop Trap: Teaching Norms Without Changing the Stage
Most psychological safety initiatives begin with a workshop. A facilitator arrives, employees sort cards labeled with phrases like “speak up” and “assume positive intent,” and everyone leaves with a shared vocabulary and a mild sense of optimism. The problem is not the vocabulary. The problem is that the workshop treats psychological safety as a communication skill, when it is more often a structural condition.
Consider a team where the manager holds a weekly status meeting in a room with a single screen, a fixed agenda, and a seating arrangement that places the most senior person at the head of the table. The manager asks for updates in a clockwise order. Junior staff speak last, after the senior staff have already framed the week’s priorities. In that room, the material setup — the table, the order, the screen — does more to shape who speaks than any amount of encouragement. A workshop on psychological safety may tell people to “make space for all voices,” but the next Monday, the same table is still there, the same order still applies, and the same three people dominate the conversation.
This is the workshop trap: it aims at beliefs while leaving the physical and procedural environment intact. The result is often a brief improvement in survey scores, followed by a slow return to baseline. Employees are not failing to learn the lesson; they are responding rationally to the structure they actually inhabit.
What Ethnography Shows About Speaking Up
In field studies of hospital teams, Edmondson found that the teams with the highest reported psychological safety also had the highest rates of reported errors. That counterintuitive finding makes sense when you look at the structure: those teams had routines for surfacing mistakes without blame, and those routines were embedded in shift handoffs, checklists, and debriefs. The safety was not a mood; it was a mechanism.
The same pattern appears in software teams, newsrooms, and factory floors. When a team has a regular, low-stakes ritual for raising problems — a daily standup that actually allows bad news, a postmortem that separates system causes from individual blame, a shared document where anyone can add an agenda item — speaking up becomes ordinary. When those rituals are absent, speaking up becomes an act of courage, and courage is a poor foundation for daily work.
Structural Levers That Actually Change the Pattern
If psychological safety is a property of structure, then the work of building it is the work of changing structure. That sounds abstract until you look at the specific artifacts of a workplace. Here are four levers that show up repeatedly in ethnographic accounts of teams that manage to shift their interaction patterns.
1. Meeting Architecture: Who Speaks, in What Order, About What
The meeting is the most visible unit of organizational culture. Its architecture — the agenda, the seating, the time allocation, the default speaker — is a material fact, not a soft skill. Teams that want more psychological safety often need to redesign the meeting before they redesign the conversation.
One practical change is to make the agenda a shared, editable document that anyone can add to before the meeting. This shifts the power to name topics away from the manager and toward the group. Another is to change the default speaking order: start with the most junior person, or with the person who has been quietest in the previous meeting, or with a written round where everyone posts one concern before anyone speaks aloud. These are not tricks. They are structural changes to the sequence of interaction, and they change who is allowed to set the frame.
Physical space matters too. A room with a round table, no head position, and chairs that can be moved changes the geometry of authority. A video call where the manager’s camera is off and the chat is open for anonymous questions changes the geometry of visibility. None of this requires a workshop; it requires a decision about how the room is arranged.
2. Performance Systems: What Gets Measured, Rewarded, and Punished
Psychological safety is often undermined not by a hostile manager but by a performance system that punishes the very behaviors the workshop encourages. If employees are rated on individual output, ranked against each other, and rewarded for being right, then admitting uncertainty is a rational risk. If the bonus depends on a project shipping on time, then raising a concern about the timeline is a threat to the bonus.
Structural change here means looking at the review rubric, the bonus formula, and the promotion criteria. Does the rubric reward asking hard questions? Does the bonus formula include a team-level component? Do promotion criteria mention mentoring, admitting mistakes, or improving a process? If not, the system is quietly teaching people to stay silent, no matter what the values poster says.
One team I observed in a field study had a simple structural rule: every project retrospective had to include at least one item that the team would do differently next time, and that item had to be written down and assigned to an owner. The rule did not require anyone to feel safe. It required the team to produce a record of learning, and over time, that record made it normal to talk about what went wrong.
3. Decision Rights: Who Can Say No, and What Happens When They Do
Psychological safety is tested most sharply at the moment of disagreement. When a junior employee questions a senior leader’s plan, what happens next? If the leader thanks the employee and adjusts the plan, the team learns that dissent is safe. If the leader explains why the plan will proceed anyway and the employee is never asked again, the team learns the opposite.
Decision rights are a structural feature. They define who has the authority to stop a project, request more data, or escalate a concern. In many organizations, those rights are concentrated at the top, and the lower levels are expected to execute. That concentration is not a personality flaw; it is a design choice. Changing it means giving teams explicit authority over certain decisions, documenting that authority, and protecting it when it is exercised.
A simple structural change is to create a “red flag” channel that bypasses the normal hierarchy: a shared inbox, a designated ombuds, or a weekly open forum where anyone can raise a concern without prior approval. The channel itself is not enough — it must be used, and its use must be visible — but the channel is a structural fact, not a feeling.
4. Physical and Digital Traces: What the Office Says About Who Matters
The material culture of work includes the objects and spaces that people encounter every day. A closed office door, a badge reader that logs entry times, a shared calendar that shows everyone’s availability, a Slack channel where only managers can post — these are all structural signals about who is trusted, who is watched, and who is allowed to speak.
In one organization I studied, the shift to psychological safety began with a small physical change: the manager moved out of a private office and into a desk in the open area. The move was not announced as a cultural initiative. It was just a change in where the manager sat. But it changed the pattern of interaction. People stopped scheduling formal meetings to raise small concerns. They started talking in the hallway, at the coffee machine, in the five minutes before a meeting. The manager heard things earlier, and the team saw that the manager was willing to be interrupted.
Digital traces matter too. A team that uses a shared document for meeting notes, where anyone can add comments, creates a visible record of who contributes. A team that uses a private channel for manager-only discussions creates a visible record of who is excluded. Neither is neutral. Both are structural choices.
Why Workshops Persist Despite the Evidence
If structural change is the real lever, why do organizations keep buying workshops? The answer is partly organizational, partly emotional. Workshops are visible, bounded, and easy to schedule. They produce a deliverable — a certificate, a slide deck, a list of action items — that can be reported up the chain. Structural change is slower, messier, and harder to photograph. It requires managers to give up some control over meeting agendas, performance rubrics, and decision rights. That is a much harder sell.
There is also a category error at work. Psychological safety is often treated as a soft skill, like communication or empathy. But the evidence from organizational research points the other way. Edmondson’s work, along with studies of high-reliability organizations, suggests that psychological safety is an emergent property of routines, roles, and artifacts. It is closer to a safety protocol than to a personality trait. You do not train people into a safety protocol; you build the protocol and then train people to use it.
This is not to say that workshops are useless. A well-designed workshop can introduce a shared vocabulary, surface hidden tensions, and build momentum for change. But a workshop that is not followed by structural change is like a fire drill that never leads to a working alarm system. The drill may be memorable, but the building is still unsafe.
A Field Note: The Team That Changed Its Meeting, Not Its Mindset
Let me offer a concrete example from a team I observed over several months. The team was part of a mid-sized product organization, and its manager was genuinely committed to psychological safety. The manager had read the books, attended the workshops, and used phrases like “no blame” and “safe to fail.” Yet the team’s survey scores on psychological safety remained stubbornly low.
The turning point came not from another workshop but from a change in the team’s weekly planning meeting. The meeting had always been run by the manager, who presented a list of priorities and asked for updates. The team members spoke only when called on, and their updates were brief and carefully worded. The manager decided to change the structure: the agenda would be built collaboratively in a shared document, the meeting would start with a silent reading of that document, and the first person to speak would be the most junior team member, not the manager.
The first meeting was awkward. The junior team member hesitated, looked at the manager, and then read from the document. The manager said nothing. The next week, the junior team member added a concern to the agenda before the meeting. The week after that, two other team members added concerns. Within two months, the team’s survey scores had improved, but more importantly, the pattern of interaction had changed. People were raising issues earlier, in public, without waiting for permission.
The manager did not become a different person. The team did not attend a new workshop. The structure changed, and the behavior followed.
What This Means for Leaders and Teams
If you are a leader who wants more psychological safety on your team, the first question is not “What training should I book?” but “What structure is currently teaching my team to stay silent?” Look at the meeting agenda. Look at the performance rubric. Look at the decision rights. Look at the physical space. Each of those is a lever, and each is easier to change than a culture.
Start with one structural change and make it visible. Change the meeting order. Add a shared agenda. Create a red flag channel. Move your desk. Then watch what happens. The team will not suddenly become candid overnight, but over weeks and months, the pattern will shift. That shift is psychological safety, not as a feeling but as a fact of the work environment.
For researchers and practitioners of workplace ethnography, this is the core insight: psychological safety is not in people’s heads. It is in the room, the schedule, the rubric, and the record. Change those, and the people will follow.
Frequently Asked Questions
Can a workshop ever improve psychological safety?
A workshop can be useful as a starting point, especially if it introduces a shared vocabulary and surfaces tensions that have been hidden. But a workshop alone rarely changes behavior because it does not change the structures that shape behavior. The most effective workshops are followed by concrete changes to meetings, performance systems, or decision rights.
What is the first structural change a team should make?
The meeting is usually the best place to start because it is visible, frequent, and relatively easy to change. A shared, editable agenda is a low-cost structural change that shifts power over what gets discussed. Changing the speaking order so that junior or quieter team members speak first is another simple change with outsized effects.
How do you measure psychological safety without relying on surveys?
Look at the traces of interaction. Who adds items to the meeting agenda? Who speaks first, and who interrupts whom? How many concerns are raised in public versus in private? How many mistakes are documented in postmortems or retrospectives? These observable patterns are more reliable than self-reported survey scores, which can be influenced by social desirability and recent events.
Does structural change work in remote or hybrid teams?
Yes, but the structures are different. In remote teams, the relevant structures include the chat platform, the video meeting norms, the shared documents, and the asynchronous communication channels. A remote team can create psychological safety by making agendas editable, using anonymous question tools, and documenting decisions in a shared space where anyone can comment.
Further Reading on This Blog
This article is part of a series on the material culture of work. Future pieces will examine the ethnography of meeting agendas, the role of physical space in authority, and the hidden signals in performance review rubrics. If you have a workplace artifact that you think reveals something about how your organization actually functions, I would like to hear about it.
