Some of the most capable people in your organization run a scan before they walk into a room.

Not of the agenda. Of the room.

Who is already in it. Where they would sit. Where the exits are. What it will cost to be fully seen today, and whether today is a day they can pay it.

You will not see the scan. It takes a second or two, and by the time the door opens the face is composed and the handshake is ready. The work happened before anyone said hello. Some of your people have already spent energy you will never get back, on a question you have never had to ask.

We have a name for what we want those rooms to have. Psychological safety. Amy Edmondson gave it to us in 1999, in a study published in Administrative Science Quarterly, and she defined it cleanly: a shared belief that a team is safe for interpersonal risk-taking.

What is easy to forget is what sent her looking. Hospitals. Earlier in that research she expected the best nursing teams to make the fewest errors. The data said the opposite. The strongest teams reported more errors, not fewer. The best teams were not making more mistakes. They were safe enough to admit the ones they made. The weaker teams were making just as many and burying them, which is the more dangerous outcome by a wide margin.

Sit with that, because it is the whole point. Psychological safety is not the absence of problems. It is the presence of enough trust to say the problem out loud.

Google confirmed it at scale years later. Project Aristotle measured everything anyone could think of about its teams: seniority, tenure, how often people socialized outside of work. The factor that predicted performance most reliably was psychological safety.

So we coach for it. We run the trainings. We circulate the lists of phrases that supposedly build it. Most of that work is sincere, and some of it helps.

Yet the whole idea rests on a word almost no one stops to examine. Safe to take a risk in the room assumes you were safe walking into the room. The definition has a precondition hiding inside it, and for a great many people that precondition is not met before the meeting begins.

Psychological safety is the second question. The first question is whether you are safe being visible at all. We have spent more than two decades answering the second while quietly assuming everyone got a yes on the first.

Not everyone did.

I call the difference the entry cost.

For people who have been treated as the exception in a room, in any of the ways that happens, the room is not neutral territory. It is terrain to be read. The scan is not anxiety and it is not weakness. It is a skill, built over years, often built young. It reads tone before content. It clocks who holds power and how they tend to use it. It runs a fast cost-benefit on decisions the rest of the room makes without thinking. How much of an accent to let through. Whether to mention the weekend honestly. Whether a story is worth telling, or safer left at the door.

That radar does not switch off when the person is competent, senior, or successful. It runs in the C-suite. It runs in the boardroom. It gets quieter and more practiced, which is not the same thing as gone.

We can measure the residue it leaves behind. Kenji Yoshino, who teaches constitutional law at NYU, built a framework for it called covering: downplaying a known part of yourself so the people around you stay comfortable. In 2023, Deloitte's DEI Institute and NYU's Meltzer Center surveyed 1,269 American workers. Sixty percent said they had covered some part of who they are at work in the previous twelve months. Among workers carrying five or more marginalized identities, it was seventy-one percent.

Then there is the number most leaders skip past. Fifty-one percent of workers with no marginalized identity said they cover as well, and fifty-four percent of white cisgender men did. If you are reading this and filing the scan under other people's problem, the data suggests you run some version of it yourself.

Covering shows up in appearance, in affiliation, in who gets mentioned and who gets left unmentioned, in the causes people quietly decline to be seen supporting. The common thread is effort spent not on the job but on managing how the job sees them. Yoshino puts the cost plainly: people end up working their identities instead of working their jobs.

Read that as a leader and the arithmetic should trouble you. You hired someone for the whole of their intelligence. A portion of it is going to the scan, the cover, the constant management of self. You are paying for the entire person and receiving the part they judged it safe to show you.

The tax does not fall evenly, and the first place you will notice it is silence.

When a room is unsafe, the first people to go quiet are usually the ones with the most to lose by speaking. That is not disengagement. It is a calculation, run faster than you can watch, by someone with years of practice running it. They have learned what speaking up cost them before, and they have decided that today is not the day to test it again.

Silence has a cost. This is what I mean by it. The perspective you most needed in the decision is the one most likely to withdraw before the decision gets made. You do not lose it in a dramatic moment. You lose it in the gap between what someone thought and what they judged it safe to say. By the time you notice the gap, the meeting is over and the call has been made without the one voice that might have changed it.

No script closes that gap. You cannot say your way into someone's safety. People who have spent their lives reading rooms can tell the difference between a leader performing safety and a leader providing it, usually inside the first few minutes.

Safety becomes real at the moment your protection of a person costs you something and you pay it anyway. When you correct the room instead of the person the room got wrong. When you take the hit in front of your own boss rather than passing it down the line. When being their leader is briefly inconvenient and you stay their leader regardless. That is the currency, and everyone who has ever had to scan a room already knows it.

I run the scan too.

I am an out gay man, married, living in the South. My husband is a Black man. When the two of us walk into a room together, we read it before we relax into it, and we read it on two frequencies at once, his and mine. Which conversations are safe to have here. Whether to hold hands, and where, and for how long. Whether the warmth we are offered is genuine or is being logged for later. None of that appears on anyone's agenda. All of it is real, every single time, and I have been doing it long enough that most people in the room never see it happen.

I tell you this so you understand that the scan is not a story about other people somewhere else. It is running in your meetings right now, behind faces as composed as mine. The people doing it are often your strongest performers, because the same vigilance that reads a room for danger also reads it for opportunity. You are getting their brilliance. You are also getting their fatigue, and you rarely get to choose which.

So before you ask your people to be brave in the room, find out what it cost some of them to enter it. Go to the person who went quiet, privately, and ask what they were about to say. Then do something with the answer where the rest of the room can see you do it. Lower the entry cost, and listen to what starts getting said.

That is the work.

In Community and Conversation, Jim.

Sources: Amy C. Edmondson, "Psychological Safety and Learning Behavior in Work Teams," Administrative Science Quarterly, 1999; Edmondson's earlier hospital error-reporting research, Journal of Applied Behavioral Science, 1996; Google re:Work, Project Aristotle; Deloitte DEI Institute and NYU School of Law Meltzer Center for Diversity, Inclusion, and Belonging, "Uncovering Culture," November 2023 (n=1,269 U.S. workers).