Researchers broadly agree that bullying exists in every human group that has ever been studied. It's something that grows naturally whenever a group forms: who's on top, who's at the bottom, who's in, who's out. The ages from eleven or twelve to fourteen or fifteen are when it's most intense, because that's the first time a person confirms who they are mainly through peers rather than family. Status suddenly matters more than anything, and the cheapest way to acquire status is to put someone else beneath you. The field begins with the Swedish psychologist Dan Olweus, who started systematic research in Scandinavia in the early seventies. At the end of 1982, three boys in northern Norway killed themselves because of bullying; the following year Norway launched a national campaign, and the intervention study Olweus ran in Bergen became the prototype for every anti-bullying program since. In 1996, Finland's Christina Salmivalli sorted the members of a classroom into six roles: bully, assistant, reinforcer, outsider, defender, victim. What she found was that most children fall into the middle roles, and whether bullying persists depends not on how vicious the bully is but on how many people are cheering and how many are watching. Bullying is a performance; without an audience it can't go on. Albert Bandura's "moral disengagement" explains how each person talks themselves into it: it's just a joke, he has problems of his own, everyone's doing it, I just didn't stop it.

For cross-national comparison, the best data we currently have is PISA. It's run by the OECD — the Organisation for Economic Co-operation and Development, headquartered in Paris, 38 member states, essentially a club of developed countries — as an international student assessment, one round every three years, testing fifteen-year-olds worldwide in reading, math, and science, with a long questionnaire about school life attached. Bullying appears as six items: in the past year, were you made fun of, left out, the subject of rumors, pushed or hit, threatened, or had things taken from you, and did it happen "a few times a month or more." The items are identical worldwide, never use the word "bullying," and ask only about actions. The "OECD average" is the mean across member states, the baseline on the chart. Mainland China isn't a member; Beijing, Shanghai, Jiangsu, and Zhejiang took part under the name "B-S-J-Z" in the 2018 and 2025 rounds and sat out 2022.

On this chart, East Asia is the quiet end. In the 2018 round, the OECD average was 22.7%; the four Chinese provinces and municipalities came in at 17.7%, Taiwan 13.2%, Hong Kong 29.3%; the Philippines topped the list at 65%, Indonesia was fifth at 41%. Japan's National Institute for Educational Policy Research put it plainly in its own summary: apart from "being made fun of by other students," Japan's victimization rates on the other five items were all below the OECD average. Under the 2015 round's "at least once in the past year" measure, Korea was the lowest of the 33 participating OECD countries, and Japan ranked 25th, also below the line. In the 2025 results released just this September 8, the OECD average is 19.6%, Japan 15.7%, Korea 13.5%. Korea's "pushed or hit" is 1.6%, second-lowest of 88 participating systems; its cyberbullying 1.2%, also second-lowest. Japan's cyberbullying is 1.1%, the lowest in the world, and 88.9% of Japanese students say they "feel they belong at this school," against an OECD average of 76.2%. The exception is Hong Kong: Esther Ho, working from the 2015 data, calculated a bullying index of 0.21 for Hong Kong, clearly above the international average; 46.7% of students had been made fun of in the past year (OECD 30.3%, Taiwan 17.4%), and 20.2% had been pushed or hit (OECD 11.9%, Taiwan 2.3%).

And yet each East Asian system's own surveys show bullying at record highs. Korea's Ministry of Education survey this year reported a victimization response rate of 2.9%, the highest in fourteen years; MEXT's recognized-case count was 769,022, an all-time high; the China Youth & Children Research Center asked middle schoolers and more than half said they'd been bullied. The difference lies in how the question is asked, and in the threshold: a Japanese child's bar for ticking "excluded" is not a Danish child's. Silent treatment and gossip are air in a Japanese classroom, not events. The item-level numbers lay this bare. In 2015, Japan's "made fun of" was 17.0% against the OECD's 10.9%; "pushed or hit" 8.9% against 4.3%; but "left out" at 4.7% and "rumors spread" at 6.1% were both below average. Japanese kids admit to being teased and shoved; they don't admit to being excluded. The only item on which the four Chinese provinces scored above average in 2018 was "property damaged or taken," 10.2% against 6.6%; the other five — mocked, isolated, rumored about, threatened — were all below. The researcher who ran that analysis said Chinese bullying leans toward "material harm" while OECD countries lean toward "harm to the person." I'd rather read it this way: Chinese kids admit that something went missing, not that something was done to them. What's distinctive about East Asia isn't how much bullying there is, it's what shape it takes, how long it lasts, and which kinds get counted as bullying at all.

On the international chart, Korea and Japan are low; on their own curves, both hit their highest point this year; and in the same year, asking the same children, PISA gets a number several times higher than the local one. These three things don't contradict each other, because the three rulers aren't measuring the same thing.

Consider the actual situation of a fifteen-year-old Korean girl in her first year of high school. In a single year she might fill out three questionnaires. In April, the Ministry of Education's survey asks her whether she "experienced school violence" last semester. She thinks about the four syllables 학교폭력 — a word that, in her world, means the school violence committee, disciplinary action, a note in her permanent record, cases that go up to the Supreme Court — and ticks "no." 99.2% of her grade answers the same way. In the fall, PISA asks whether in the past year other students made fun of her, deliberately left her out, spread rumors about her, pushed her, threatened her, took her things, a few times a month or more, and she ticks "made fun of" and "left out." 13.5% of the country answers that way. She didn't lie either time. The first questionnaire asked "are you a person who has been subjected to school violence"; the second asked "did these things happen." In Korea, the proportion of children who will admit the latter is seventeen times the proportion who will admit the former.

Japan runs the other way. MEXT's 769,022 cases were filed by schools, not by children. Divide by the roughly twelve million-plus primary and secondary students nationwide and you get 61.3 per thousand — about 6% of children "recognized" by their schools as having been bullied in a year. PISA asks Japanese fifteen-year-olds themselves, a few times a month or more, and 15.7% say yes. So Japan's situation is: children report more than twice what schools recognize, and the 6% schools recognize is already an all-time high. MEXT itself concedes the number measures the schools' eyes, not the bullying — a sixfold gap between prefectures can't be a gap between children. China is a third case: no school count, no government survey of children, only researchers' samples. The China Youth & Children Research Center asks middle schoolers "have you ever been bullied," and more than half say yes; PISA asks fifteen-year-olds in the four provinces "a few times a month or more," and gets 17.7%. The first is "ever," the second is "often"; the first uses the word "bullying," the second lists six actions.

PISA only asks fifteen-year-olds — first year of high school in Korea and Japan, first year of high school or last year of middle school in China. All three systems' own data say bullying peaks in primary and middle school: Korea's primary rate is 5.7%, middle 2.3%, high 0.8%, falling straight down; two-thirds of violent incidents in Japan happen in primary school; Gansu's fourth-graders, 62%. What PISA photographs is the beach after the tide has gone out. A child who was ignored by the entire class for a year in fourth grade may well be in none of the numbers by the time they fill out PISA at fifteen. One of PISA's six items is "other students deliberately left me out." A fifteen-year-old in Denmark or Canada reads that and thinks of a party they weren't invited to; a Japanese kid reads it with 仲間はずれ and 無視 as the frame of reference — weeks in which nobody speaks to you. When the "event" in your frame of reference is that severe, everyday small exclusions don't count, and the box stays empty. That's why Japan is below the OECD average on "left out" and above it on "made fun of" and "pushed" — not because Japanese kids don't exclude people, but because in Japan, how severe exclusion has to be before it deserves the name is a higher bar than elsewhere. The Chinese result, with "property" as the only item above average, follows the same logic: a thing going missing is tangible and reportable; "something being done to a person" doesn't yet have a name many Chinese kids can tick.

None of these numbers measures bullying. Each measures the distance between a child and a word: whether they're willing, able, and likely to call what happened to them by that word. Korea made the word very heavy, so children stand far from it; Japan handed the word to schools, so schools decide the distance on children's behalf; China never institutionalized the word, so each researcher sets the distance themselves. A good part of East Asia's "low" on the international chart is the length of that distance.

The Japanese sociologist Naito Asao's Ijime no Kōzō (The Structure of Bullying, 2009) is built around the concept of "intermediate-group totalitarianism": Japanese schools slice children by age into a class and lock them in the same room all day — meals, cleaning, clubs, group activities, all together, with no freedom to choose companions and no possibility of exit. In this structure the group itself becomes the source of morality; whatever "everyone" thinks is right, being cast out of "everyone" is the heaviest punishment, and casting someone out is the cheapest structure" — bully, victim, audience, bystanders — but he put the weight on the bystanders: in Japan the proportion of bystanders rises with age and is a majority by middle school, and in Japanese culture standing by isn't neutral, it's tacit consent.

This explanation holds for China and Korea more or less intact, with each adding its own weight. The common base across all three is: fixed classes, ten-plus hours a day, admission to the next school as the sole goal, zero-sum competition among children, and teacher authority resting on "nothing happening." China adds scale and boarding — in the Gansu data, boarders score higher, and so do classes with worse teacher-student ratios — and it adds the left-behind child: a kid whose mother is away has no one to tell at home, and therefore no one to speak for them at school. Korea adds hierarchy: the seniority sequence of sunbae and hubae, the class consciousness of a chaebol society, and a public language that converts everything into "fairness," which is why Chung Sun-sin's son, bullying a classmate, reached for the other boy's "Jeju origins" and "family political background." That's not random malice; it's a child replicating the ranking system of the society he lives in. Japan adds "air": never saying it out loud, replacing beatings with silence and exclusion, which is why "taunting and teasing" makes up sixty percent of Japanese bullying and physical violence only a few percent, and why it stretches on so long.

Erving Goffman, the Canadian-American sociologist, published Asylums in 1961, based on a year of fieldwork in a Washington psychiatric hospital. In it he proposed the concept of the "total institution." The definition is simple: a place where a group of people eat, sleep, work, and play with their whole lives under one roof, one authority, one timetable, cut off from the outside world. His examples were psychiatric hospitals, prisons, army barracks, monasteries, boarding schools, ocean-going ships. In normal society those three things — sleeping, playing, working — happen in different places, with different people, under different authorities; the total institution merges them. He observed three things. First, on entering, a person undergoes a "mortification of the self": clothes replaced, name changed or swapped for a number, private possessions and private space gone, and from then on who you are is for the institution to say. Second, wherever the formal rules don't reach, an "underlife" grows — informal hierarchies, trades, punishments, slang — and the institution knows this perfectly well, because that underlife helps it keep order. Third, the managed develop adaptation strategies: total withdrawal, open rebellion, "colonization," in which the system becomes the whole world and one stops wanting to leave, or outward compliance with inner reservation.

Goffman didn't count ordinary schools as total institutions; boarding schools were on his list. East Asian day schools, strictly speaking, aren't total institutions — children go home at night. But several things bring them close: ten-plus hours a day, fixed classes, no freedom to choose companions, meals and cleaning and clubs all within the same group, a single standard of evaluation, and exit only through the entrance exam. Naito's "intermediate-group totalitarianism" is essentially Goffman's framework moved into the Japanese classroom; in his words, the school puts children into "a small society you can't escape." China's boarding schools, and the higher victimization rate among boarders in the Gansu data, come even closer to Goffman's original. When a group of people are locked in together with no exit and formal rules that only reach the surface, informal hierarchies and punishments will grow on their own, whether the people inside are fourteen or twenty-four. This isn't a theory about bad people. It's a theory about space. 📚