Theme 28 of 43

School Violence, Educational Security & Learning Environment

Students who feel unsafe at school score 15% lower in math — bullying is the most common form of violence inside Brazilian schools, and also the least reported.

The School Violence Atlas, maintained by the Brazilian Public Security Forum, classifies incidents recorded inside schools by type — bullying, physical assault, weapons possession, theft, drugs, vandalism — and cross-references them with territory and school network. SAEB, Inep's national learning assessment, asks students directly whether they feel safe at school and asks teachers whether they address bullying in class, making it possible to link that perception to measured performance in math and Portuguese. The School Census details each school's physical infrastructure — walls, fencing, lighting — and its ties to public security forces, while crime records from Rio de Janeiro's ISP and violence notifications from SINAN, the Health Ministry's system, supply the community backdrop that spills over the school walls. Together, these records show that schools are not a protected bubble insulated from the violence surrounding their neighborhood — they are, in large part, its mirror.

Bullying is the most common form of violence — and the least reported

Among all incidents recorded in Brazilian schools, bullying leads by a wide margin, ahead of physical assault, theft, weapons possession, and drugs combined. It's also the type of violence most normalized within the school environment: because it leaves no immediate physical mark and doesn't require police intervention, it tends to be treated as part of the routine among students — which in practice means systematic underreporting. The real number of cases is, in all likelihood, still higher than the already substantial recorded share.

Bullying is the most common form of violence inside Brazilian schools and, at the same time, the most invisible — normalized to the point of escaping official reporting.

State schools concentrate more violence

The school network a student attends already predicts, to a significant degree, the type and intensity of violence they're exposed to. State schools register higher incident rates, concentrated in physical assault and bullying, while the private network keeps low rates — though still subject to cyberbullying, a form of violence that doesn't depend on a school's physical space to occur. The difference isn't born inside the classroom: it reflects the violence of the territory each network sits in.

State schools concentrate more violence than any other network — a direct reflection of the community violence surrounding them, not an internal pedagogical failure.

Fear inside the school steals performance

Students who tell SAEB they don't feel safe at their own school score, on average, 15% lower in math than their peers — an effect comparable in magnitude to other major educational inequalities already documented in Brazil. Schools without a security guard record double the incidents, and those located in drug-trafficking areas register triple the assaults. Fear isn't a side effect of school violence: it's the mechanism through which that violence converts directly into lost learning.

Feeling unsafe at school cuts math performance by 15% — fear has a direct, measurable effect on learning.

Militarized schools: mixed results

The number of schools with a formal tie to public security forces has been growing across the country, but this model's effect on school violence is far from settled in the data. Some incidents fall, but others persist or worsen, and the correlation between militarization and academic performance turns out negative — a sign that policing inside schools doesn't substitute for pedagogical and psychosocial prevention work.

Militarized schools show mixed results — they reduce some types of violence, but the effect on academic performance is negative.

The regional map of school violence

The Southeast and North appear at the top of the school-violence-rate ranking, but for different reasons: in the Southeast, higher density of records may in fact reflect more incidents; in the North, it's plausible that the recorded rate hides greater underreporting, given the region's thinner network of reporting systems. The predominant type of violence also varies — bullying and drugs in the Southeast, physical assault in the North — suggesting each territory produces its own signature of school violence.

The Southeast records more school violence — but the North may simply be underreporting the problem, not experiencing less of it.

Explanatory hypotheses

School violence reflects the reproduction of social inequality within school walls: public schools located in vulnerable areas concentrate more incidents because they also concentrate more exposure to community violence. The connection to territory shows that schools function as a mirror of the surrounding neighborhood, not a space isolated from it. And the correlation between perceived insecurity and declining performance confirms that a sense of safety is a prerequisite for learning — without it, pedagogical effort loses effectiveness even when teaching itself is high quality.

Policy implications

Conflict-mediation programs among students could reduce bullying without resorting to militarizing the school environment. Full-time schools could productively occupy vulnerable teenagers, reducing exposure to the highest-risk hours. Strengthening the family-school bond could contain domestic violence that spills over into the classroom. And having school psychologists present at every school could enable early identification and intervention, before violence turns into dropout or a permanent decline in performance.