School-Related Violence Research: Where Next?

Violence in and around schools adversely affects children’s well-being. It also affects their ability to learn, making it a central challenge for education systems. A growing body of research on school violence—spanning public health, education, and economics—demonstrates both the gravity of the problem and a set of effective interventions governments can act and build on now.

But important knowledge gaps remain. This blog post builds on discussions at a Research Workshop on School Violence Prevention and Response earlier this year—for which the five of us served on the scientific committee—along with our own knowledge of existing research to highlight six priority areas for future research.

These proposed priorities point to where the evidence is more limited. They include (1) a deeper understanding of what drives the effectiveness of specific interventions and policies, (2) more evidence on how to strengthen violence prevention and response systems, (3) adult-to-child violence (not just peer-to-peer), (4) shifting the social norms that undergird violence, (5) the links between violence and learning at school, and (6) the longer term impacts of interventions to reduce violence. At the end, we highlight a series of issues that cut across all research on school-related violence. Filling these gaps will strengthen the evidence base needed to inform the design of policies and programs that are effective, sustainable, and scalable.

Six priorities for future research

1. Mechanisms of change: Better understand what makes interventions work, for whom, and in which contexts.

Evidence on interventions aimed at reducing school-related violence shows that much violence is preventable. But most effective interventions are packages of activities, so we need to explore in more depth which design and implementation features and intervention components are most effective, for whom, and in which contexts.

To achieve this, we need more studies that collect and use implementation data—combining qualitative and quantitative approaches—to unpack what is working, what isn’t, and why. This includes using data to explore whether an intervention design was appropriate, whether it was implemented as intended (fidelity), and how beneficiaries responded (take-up). For example, in Zambia, implementation data helped explain why a whole-school approach intervention had no impact on students’ knowledge and experiences of violence. The analysis showed that implementation challenges—such as reliance on voluntary teacher mentors, a lengthy curriculum, and inconsistent delivery—undermined fidelity, and that low take-up among teachers and parents due to specific context barriers potentially led to a lack of impact on the measured outcomes.

Because most school-violence prevention interventions are multi-component and often evaluated in randomized controlled trials (RCTs) with a single treatment arm, it remains unclear which “ingredients” matter most. Multi-arm RCTs are a useful approach for disentangling the relative contribution of different components and identifying which actors to target in resource-constrained settings and at scale.

Even as we generate more evidence on the impact of interventions, more research comparing the effectiveness of existing interventions will be useful. To advance on this, standardising violence measures and cost reporting across studies would enable meaningful comparisons of what is more cost-effective.

2. Prevention and response: Increase efforts to understand what effective response systems look like, while continuing to advance research on how to prevent violence.

A key gap in the literature relates to understanding how to strengthen education system responses to violence. This includes school reporting systems, grievance and response mechanisms, and referral pathways that are integrated with the broader child protection systems. A systematic review (work in progress) on school-related violence interventions across high-, middle-, and low-income countries shows that of more than 40 interventions included in the review, nearly all include a prevention component, while one in eight include a component related to violence response. Yet improving response shows promise in reducing violence in and around schools, increasing school enrolment, and reducing student transfers.

Prevention remains essential, as it aims to address the root causes of violence, and ongoing investments in both research and programming in this area must continue! At the same time, as we continue to invest in prevention, there are at least three reasons why expanding research on how to strengthen response systems can’t be ignored if we want to change schools’ operational culture. First, weak or absent response mechanisms can undermine prevention efforts: when school environments fail to hold perpetrators accountable, violence is more likely to happen again, and children are more likely to drop out of school. Second, response systems are particularly important for addressing adult-to-child violence, where power imbalances limit children’s ability to seek help. Third, effective response systems not only address school-related violence but also enable schools to fulfil their mandate to identify and report suspected cases of violence in the home environment.

3. Adult-to-child violence: Increase research on adult-to-child violence, particularly sexual violence, which remains under-researched compared with peer-to-peer violence.

We need more research on how to address adult-to-child forms of violence in, around, and on the way to school. Evidence exploring how to address corporal punishment shows that whole school approaches, aimed at shifting the school operational culture, have been effective in reducing teacher physical violence in Uganda. Research in Mozambique also shows that strengthening reporting mechanisms and subsequent school responses to violence can reduce sexual violence perpetrated by teachers.

Expanding this evidence base is essential, particularly given the power dynamics that underpin abuses such as cases of “sex for grades” and that discourage students from speaking up. These challenges are often worsened by the prevalence of weak institutional accountability, where, in practice, reports of teacher-perpetrated violence do not consistently lead to action.

4. Social norms: Deepen our understanding of how attitudes and behaviours change in order to shift social norms.

More evidence on what and how to effectively shift norms, attitudes, and behaviours around violence is essential to address risk factors in a sustained way. Corporal punishment remains lawful in 62 countries, and even where it has been banned, evidence suggests that bans alone are not enough to change practices and attitudes. In Uganda, for example, although tolerance has declined, corporal punishment is still widely perceived as acceptable. This highlights the need for more evidence on how laws interact with existing norms, and how monitoring and enforcement can strengthen compliance.

More research is needed on how to support transitions toward bans that are both adopted and effectively implemented. Evidence from efforts to address practices such as female genital mutilation suggests that “stepping-stone” approaches—introducing incremental changes on the path to fully replacing harmful norms—can be effective. In the context of corporal punishment, promoting positive discipline or fostering community dialogue may help shift norms before bans are enacted.

We also need to understand norms that shape reporting and experiences of violence, including those that minimize boys’ experiences or normalize silence. Adopting an intersectional lens is critical, as norms—and how they change—vary across gender, disability, socioeconomic status, and location.

5. Violence and learning: Demonstrate more clearly the nuanced relationship between violence and learning.

Existing evidence shows a strong association between school-related violence (mainly bullying) and learning outcomes. However, stronger evidence of the causal nature of this relationship, along with a clearer understanding of the mechanisms through which violence affects learning, can motivate education leaders and funders to work more on school safety. We highlight two key research gaps.

First, we need better data on the causal link between school-related violence and learning. Establishing causality is challenging because exposure to violence is not random. Panel data and quasi-experimental methods can help account for issues related to reverse causality and selection bias. For example, in Chile, researchers used longitudinal data and a difference-in-differences design to show that increases in reported physical and emotional violence from peers and teachers increased student absenteeism, increased grade retention, and reduced test scores. Studies using similar approaches across different contexts could help further explore this causal relationship.

Second, the flip side of the last question is understanding the impact of school-based violence prevention interventions on learning and through which mechanisms (for example, improved attendance or school climate). Evidence on this is still limited and mixed, with a few studies finding some or no short-term effects on learning. (Even if reducing violence doesn’t boost learning, it’s still worth doing!) Future evaluations should explicitly include higher quality measurement of learning outcomes, analyse the conditions that could limit or enhance the effect of reducing violence, and be adequately powered to detect changes in academic performance.

Beyond narrow definitions of learning (academic test scores), additional outcomes merit attention. These include children’s broader well-being, including their mental health.

6. Long-term impacts: Learn more about whether the effects of interventions, policies, and laws are sustained over time.

Most of the evidence studying the impacts of interventions, policies, and laws is short-term. This leaves important gaps in understanding whether impacts persist, evolve, or dissipate over time. We recommend investing in evidence that can expand understanding of whether intervention impacts are sustained over time (for example, whether reductions in violence at school persist beyond project completion and whether the positive impacts of that reduction extend beyond children’s school careers). Some effects may only emerge later, such as improvements in learning that only follow sustained improvements in school climate. Similarly, it is also important to understand how laws and policies shape norms, attitudes, and behaviours around violence over longer time horizons, including potential intergenerational changes, as these dynamics are essential to inform how to design policies that produce sustainable change.

Moving forward

We present six key areas for future research in a growing body of work with important gaps. Our proposed priorities are not exhaustive. Other researchers in this space will surely highlight other crucial areas: for example, UNESCO has a forthcoming document on priority setting in school-related violence research. Beyond our main points above, researchers, policymakers, and funders must keep in mind the importance of:

• Carefully considering data collection decisions, prioritising survey methods that minimize misreporting and the use of standardized measures where possible.
• Actively fostering and funding cross-disciplinary collaboration, thus avoiding siloed approaches.
• Recognising that the challenges of researching school violence call for creative, rigorous research designs—qualitative, quantitative, and mixed methods—centring the voices of young people.
• Including cost data in impact evaluations to better assess cost-effectiveness.
• Ensuring rigorous safeguarding practices are embedded throughout the research process.
• Linking evidence generation to policymaking by engaging governments early, co-designing interventions with grassroots organizations, students, and parents, and following up for policy impact beyond publishing a research paper.
• Examining how technology both facilitates violence and may be an instrument to prevent and report violence.

As researchers, it can be hard to balance the joint facts that there is clear evidence that policymakers can act on to reduce violence and that there is much left to learn. We encourage bold action and clear investments in learning along the way.

Many thanks to Xavier Hospital, Elsa Burzynski, and Sally Zweimueller for their comments and feedback.