Restorative Circles In Schools — What The Data Shows

The logic of punitive school discipline is simple enough to state in a sentence: make the consequence bad enough that the behavior won't be repeated. This is deterrence theory applied to children. It sounds reasonable. It is, at the level of evidence, almost entirely unsupported.

The research on suspension, expulsion, and zero-tolerance policy has been accumulating for two decades, and the picture it paints is consistent enough to call a verdict. The American Psychological Association Zero Tolerance Task Force, in a landmark 2008 review, examined the evidence and found that zero-tolerance policies failed on their own terms. They did not produce safer schools. They did not reduce behavioral problems over time. They did not improve academic outcomes. What they did produce was higher suspension rates and, downstream from that, higher dropout rates and higher rates of contact with the criminal justice system.

A 2011 study in Texas by Mike Shollenberger — tracking 900,000 students over six years — found that students who were suspended or expelled were three times more likely to have contact with the juvenile justice system within the following year. Three times. That's not a marginal correlation. That's a mechanism. Removal from school increases, not decreases, the probability of future harm.

The mechanism isn't mysterious. When a student is suspended, several things happen simultaneously: they fall behind academically, making school less rewarding and more aversive; they lose connection to the social and institutional fabric of school, which is often one of the few sources of structured belonging in their lives; and they spend unsupervised time in environments that may actively reinforce the behaviors the school was trying to address. The suspension accomplishes what the school intends — removal — while accelerating the conditions that produce the behavior it was trying to stop.

And the distribution of this harm is not random.

The data on racial disparities in school discipline is not new, but it has become impossible to ignore. The U.S. Department of Education's Civil Rights Data Collection has documented it repeatedly and consistently: Black students are suspended at 3.8 times the rate of white students. That disparity shows up in preschool — where Black children represent 18 percent of enrollment but 48 percent of suspensions. It shows up in elementary school. It tracks through secondary school. And it persists even when researchers control for poverty, disability status, school location, and infraction type.

This is not a matter of Black students committing more infractions. Multiple studies using observational and audit methodologies have documented that the same behavior — the same verbal altercation, the same level of disruption — results in harsher disciplinary responses when the student is Black. A 2019 study in the journal Psychological Science found that teachers who were shown identical behavioral descriptions attributed more "troublemaking" to Black students and recommended more severe punishments, and that this effect intensified when the teachers had just been reminded of prior incidents — that is, when they were primed to see a pattern. The racial discipline gap is, in significant part, a product of differential perception and differential response to identical behavior.

The downstream effects compound. Black students who are suspended are more likely to drop out, more likely to be arrested, more likely to end up in the criminal justice system. The school-to-prison pipeline is not a metaphor. It is a set of documented statistical relationships connecting school discipline practices to mass incarceration, operating with particular force on Black boys and, increasingly, Black girls.

Any discipline system that produces these outcomes is not, in any meaningful sense, working. It is producing harm in the name of preventing it.

Restorative justice as a formal concept has roots in multiple traditions. Maori communities in New Zealand used communal conferencing practices for harm resolution long before Western legal systems took notice. Indigenous practices across North America operated on principles of community repair rather than individual punishment. Quaker peacemaking traditions contributed to the modern restorative justice movement in the 1970s and 1980s, as did the work of criminologists like Howard Zehr, whose 1990 book Changing Lenses reframed justice not as punishment of an offender by the state but as repair of relationships between people.

The core distinction Zehr drew was this: retributive justice asks "What rule was broken? Who broke it? What punishment do they deserve?" Restorative justice asks "Who was harmed? What do they need? Whose obligation is it to meet those needs, and how?" These are genuinely different questions, and they produce genuinely different processes and outcomes.

Applied to schools, restorative practices take several forms. Affirmative circles build community proactively — classrooms that use regular check-in circles to build relational trust before any incident occurs. Responsive circles address harm after it has happened, bringing together the harm-doer, the harmed, and affected community members. Re-entry circles support students returning from suspension or other absence, helping reintegrate them into the community rather than dropping them back in without acknowledgment of what happened. Each of these formats is grounded in the same underlying principle: harm is a rupture in relationship, and repair requires relationship.

Oakland Unified School District is the most-cited case study in school restorative practice, for good reason. Starting in 2006 with a pilot in four schools and expanding district-wide through a dedicated Office of Restorative Practice, Oakland committed to a structural shift. By 2011, suspension rates had fallen by 44 percent. By 2016, overall suspensions were down 87 percent from 2006 baseline numbers. Chronic absenteeism declined. The racial suspension gap narrowed meaningfully, though it did not close entirely. Teacher survey data showed substantial improvements in school climate scores — teachers reporting they felt supported, students reporting they felt the school was a place where they belonged.

The Oakland data is sometimes criticized for not being a clean controlled experiment — which is fair. District-wide implementation without a control group makes it hard to isolate the specific effect of restorative practice versus other concurrent changes. That critique is methodologically legitimate. But it is also the case that the trend lines are large and consistent, and they moved in the right direction at the right time.

Denver's restorative pilot, implemented in Denver Public Schools starting around 2013 with external evaluation by the Denver School-Based Restorative Practices Partnership, produced more tightly studied results. The evaluation documented significant reductions in disciplinary referrals, out-of-school suspensions, and racial disparities in discipline in schools that implemented restorative practices with fidelity. Critically, it tracked recidivism — what happened to students who went through restorative processes versus those who received traditional punishments for comparable infractions. Students who went through circles were substantially less likely to be referred for discipline again within the same school year. The process changed behavior. The punishment did not.

A 2019 randomized controlled trial conducted in Pittsburgh Public Schools — the most methodologically rigorous study to date — found that schools randomly assigned to receive restorative practice training and implementation support had significantly lower rates of suspension than control schools by the end of the implementation period, with the largest effects on Black students. The suspension gap between Black and white students was substantially smaller in restorative schools.

The design of a restorative circle is not incidental. It is the mechanism.

A talking piece is central. Whoever holds it speaks; everyone else listens. This simple protocol does something to power: it distributes the right to speak rather than concentrating it in the facilitator or the most dominant person in the room. Students who have never been listened to — whose presence in school has been primarily experienced as a problem to manage — hold the talking piece and speak, and are heard.

The facilitator uses open-ended questions designed to elicit narrative rather than defense. Not "Why did you do that?" — which primes a student to justify or minimize — but "What were you thinking about when this happened?" and "Who else was affected, and how?" and "What would it take to make this right?" These questions require the person who caused harm to move cognitively and emotionally outward — to imagine the impact of their actions on another human being, rather than focusing on managing their own consequences.

Research on perspective-taking and moral development supports why this matters. Zahn-Waxler and colleagues documented in the 1970s and 1980s that the capacity for moral concern is not fixed; it is activated or suppressed by social context. Children raised in environments where they are asked to consider the feelings of others develop stronger and more durable prosocial behavior. Restorative circles build exactly that capacity — systematically, repeatedly, in the context of real situations with real stakes.

The agreement at the end of a circle is not a sentence handed down. It is a negotiated commitment. The student who caused harm participates in defining what repair looks like. This matters for buy-in — people follow through on commitments they helped construct. But it also matters for identity. A punishment defines you as someone who broke a rule. A restorative agreement defines you as someone who can repair what you broke. That's a different narrative about who you are, and narratives about identity have predictive power over behavior.

The most common objection to restorative practices from parents, teachers, and school administrators is some version of: "That's too easy on kids who did something wrong. There have to be real consequences."

This objection reveals a particular theory of what consequences are for: pain. If the consequence doesn't hurt enough, it doesn't count. This is a deeply held cultural assumption in societies shaped by punitive legal traditions, and it is worth examining directly.

Sitting in an in-school suspension room for three days is not particularly demanding. You don't have to think about anything. You don't have to face anyone. You don't have to account for anything. You serve the time and you're done.

Sitting in a circle with the person you hurt, while a trained adult facilitates them telling you what your actions did to their life, and then building a concrete plan for how you will repair what you broke — that requires something from you. It requires presence. It requires accountability in the literal sense: the capacity to give an account of yourself, to stand in front of another human being and be seen. Many students report that the circle was the hardest thing they ever did in school. That is not because it was gentle. It is because it was real.

The question is not whether consequences should exist. The question is whether the consequence does the work we say we want it to do — changing behavior, restoring safety, repairing harm. On all three measures, restorative practices outperform punitive ones. They are harder to implement, require more training and time, and demand more from adults and students alike. They are not easy. They are more effective.

Children learn how conflict is handled, how harm is addressed, and who deserves repair through their direct experience of institutional life. For most children, the primary institution of their childhood is school. The model they absorb there — consciously and unconsciously — shapes how they will navigate conflict for the rest of their lives, and how they will build and participate in institutions as adults.

Schools that use suspension as a primary discipline tool are teaching children: when harm happens, we remove the problem person. This does not build the capacity to repair. It builds the expectation that people who cause harm are removed, which transfers into community life as the expectation that punishment is the only available response to wrongdoing.

Schools that use restorative circles are teaching children: when harm happens, we bring people together to repair it. Harm is addressed, not avoided. Accountability is communal, not purely individual. People are capable of making things right. That lesson — absorbed in elementary school, reinforced through secondary school, practiced in real contexts — builds a different kind of adult. An adult who knows that hard conversations are survivable, that repair is possible, that community holds rather than discards.

Scale that across a generation of children. Imagine schools everywhere, in every country, practicing this. The children who grow up inside that system carry it with them into every institution they build and inhabit: their workplaces, their local governments, their families, their communities. The circle teaches something that punitive discipline cannot: that we are accountable to each other, not just to rules.

That is Law 1 made structural. That is We Are Human built into the architecture of how the next generation learns to be in the world.

For educators: Introduce a weekly community circle in your classroom — not as a response to incidents but as a proactive relationship-building practice. Use a simple talking piece and one question: "What's something that's been on your mind this week?" Do this for six weeks before you need it for anything disciplinary.

For parents and community members: Attend a school board meeting where discipline policy is discussed. Ask: What are our suspension rates? What is the racial breakdown? What is the recidivism rate for students who are suspended? Request data. Propose a restorative pilot.

For anyone: The next time there is a conflict in your own life — at work, in family, in community — try the restorative question sequence before defaulting to avoidance or punishment: Who was harmed? What do they need? What would repair look like? Notice what changes when you ask.

- American Psychological Association Zero Tolerance Task Force. "Are Zero Tolerance Policies Effective in the Schools?" American Psychologist, 2008. - Shollenberger, T.L. "Racial Disparities in School Suspension and Subsequent Outcomes." Closing the School Discipline Gap, 2015. - Anyon, Y. et al. "The Persistent Effect of Race and the Promise of Alternatives to Suspension in School Discipline Outcomes." Children and Youth Services Review, 2014. - Davis, J.E. and Farbman, D. "Restorative Practices in Denver Public Schools." Denver School-Based Restorative Practices Partnership Evaluation, 2018. - Augustine, C.H. et al. "Can Restorative Practices Improve School Climate and Curb Suspensions?" RAND Corporation, 2018. (Pittsburgh RCT) - Zehr, Howard. Changing Lenses: A New Focus for Crime and Justice. Herald Press, 1990. - U.S. Department of Education Office for Civil Rights. "2017-18 Civil Rights Data Collection: School Climate and Safety," 2021.