Restorative justice is sometimes dismissed as soft, as a way of avoiding accountability in the name of healing. This characterization inverts the actual experience of people who go through restorative processes. The evidence suggests that restorative justice produces more accountability, not less, and more durable behavioral change than conventional prosecution.
The core mechanism of restorative justice is a facilitated encounter between the person who caused harm, the person or people harmed, and members of the broader community affected. The purpose is not to determine guilt, which may already be established, but to explore the impact of the harm, agree on what accountability and repair look like, and create conditions for genuine resolution. Participation is voluntary for all parties.
What this looks like in practice varies considerably across programs. In school-based restorative practices, a student who has caused harm might participate in a facilitated circle with the students they affected, their families, and school staff. The conversation surfaces the actual impact of the conduct, creates space for genuine apology, and produces an agreement about what repair looks like. In victim-offender mediation programs, adults convicted of property crimes or some assault offenses can participate in mediated dialogue with victims who request it. In serious crime contexts, survivor-centered programs like the Resolve to Stop the Violence Project in San Francisco have used restorative principles in prison settings to address cycles of violence.
The research on restorative justice outcomes is more rigorous than critics often acknowledge. Lawrence Sherman and Heather Strang at Cambridge University conducted a meta-analysis of randomized controlled trials of restorative justice and found consistent reductions in reoffending across offense types, with stronger effects in cases involving personal violence than property crime. Victims who participated in restorative processes reported higher satisfaction than those who went through traditional prosecution, higher rates of receiving apologies, higher rates of agreement compliance, and lower levels of post-traumatic stress symptoms.
The satisfaction differential is notable. Traditional prosecution centers the state as the primary victim of crime, reducing the actual harmed party to a witness. Restorative processes center the person harmed in determining what accountability and repair mean. Victims in these programs are not passive recipients of a verdict. They are active participants in the resolution. For many people who have been harmed, that participation is itself part of healing in a way that watching a prison sentence be handed down is not.
The New Zealand family group conferencing model, which became the default response to youth offenses there in the early 1990s, provides the most comprehensive national evidence. Youth reoffending rates dropped substantially over the years following its implementation. Victim satisfaction rates are consistently high. The program has been maintained across multiple changes in government, an indication that the evidence base is durable enough to survive political cycles.
Restorative justice is not appropriate for every case. It requires voluntary participation from all parties, and some victims do not want to engage. It is not designed to respond to serious violent offenses without substantial modification and additional safeguards. But for the many cases where it is appropriate, the evidence for its effectiveness is stronger than for most conventional interventions. The question is not whether it works. It is whether we are willing to design systems around what works.