The opening chapter of the ArtPartner guide is about a decision — not a policy, not a poster campaign, not an HR procedure. A decision made by one adult, usually within ninety seconds of seeing something they wish they hadn’t seen. That decision determines almost everything that follows.

We didn’t choose to start there. The research did.

The number that hasn’t moved — and what that tells us

The CDC’s Youth Risk Behavior Survey is the longest-running snapshot of adolescent wellbeing in the United States. Its 2023 results report that 19% of US high school students were bullied at school in the past year, up from 15% in 2021. Across two decades of awareness weeks, t-shirts, and pledge cards, the number has hovered in the same band — high teens, sometimes nudging twenty.

That stubbornness is the point. It tells us the dominant strategy — tell people bullying is wrong — is not the lever. People already know. Sixth-graders can recite it. The 19% don’t persist because of a knowledge gap. They persist because of a behaviour gap, in adults.

What Olweus actually found in Norway

The Olweus Bullying Prevention Program is the most studied programme of its kind in the world. Dan Olweus began the work in Bergen in the late 1970s, and after three Norwegian boys took their own lives in 1982 and 1983, his nationwide intervention study became the template every later programme has tried to beat. Forty-plus years later, almost nothing has.

The programme has four levels — school-wide, classroom, individual, community — and the manuals are detailed. But when researchers ask which component carries the effect, the answer is consistent across the literature. It isn’t the assemblies. It isn’t the curriculum hours. It isn’t the parent letters. It is the change in the bystander norm: the moment when the adult or the peer in the room stops watching and intervenes, calmly and without escalation, by name.

StopBullying.gov, the federal site coordinated by HHS, leads with the same idea. Its prevention guidance is built around adult presence and predictable, immediate response — not punishment ladders, which the research consistently shows backfire when used as a first move.

Maple Ridge Middle — what one decision looked like

The chapter walks the reader through a scene from a real US middle school we’ll call Maple Ridge, in a small Ohio district. (The school’s name and the names of the students have been changed; the principal reviewed the chapter before publication.) Lunch period, eight minutes in. A seventh-grader the staff knew was struggling — we’ll call him Marcus — was being filmed by three classmates while a fourth tipped his tray. The cafeteria monitor, a long-tenured staff member, was eight metres away.

Two things happened in the next ninety seconds. First, the monitor walked over — not at speed, not raising her voice. She placed a hand on the back of an empty chair next to Marcus and said, by name, to each of the four other students: “Phones down. Trays back. You’re sitting with me for the rest of lunch.” Second, and this is the part the chapter circles back to, she said the same line to two students at the next table over — the ones who had laughed but not filmed. You’re sitting with me for the rest of lunch.

That second sentence is the decision. Most adults stop at the four. Most policies are written for the four. The Olweus literature, the CDC’s school-climate work, and the practitioner data we’ve drawn on for the guide all point at the same thing: the cost of bullying is set by the audience, not the actor. When the audience pays a small, predictable, immediate cost — sit with me, walk with me, repeat back to me what you saw — the next incident drops by an order of magnitude. Maple Ridge tracked a 70% drop in reported lunch-period incidents over the following semester. The cafeteria didn’t get any new posters.

What this means for an office, not a school

The ArtPartner programme is built for our Paris office, not for a middle school in Ohio. The legal frame is different — harcèlement moral under the French Labour Code is not the same construct as US peer victimisation. The age and power dynamics are different. But the underlying mechanic, the one Olweus put a name to in 1983, is identical. A team meeting in which one colleague is consistently spoken over. A Slack thread in which one designer’s ideas are quoted back without credit. A pattern of who gets invited to the after-work drink and who never does.

In every case, there is an audience. In every case, the audience pays no cost. And in every case, the most senior person within line of sight has ninety seconds to make a decision. That decision is the chapter.

The cost of bullying is set by the audience, not the actor — and the audience is almost always us.

Chapter 1 is free because it is the chapter that does the most work. The rest of the guide — the workshop, the handbook, the e-learning modules, the compliance templates — exists to make that ninety-second decision easier, more practised, more predictable, for everyone in the office. If the chapter lands, the rest is rehearsal.