The research team used a random-effects model, the standard statistical framework for combining results across studies that differ in design, samples, and settings. This approach treats the true effect of interventions as varying from study to study rather than assuming a single universal value, which is essential when pooling trials that range from kindergarten aggression-prevention lessons to web-based cyberbullying programs for adolescents. The pooled results showed that intervention programs were associated with a reduction in bullying perpetration with an effect size of 0.78, and a reduction in victimization with an effect size of 0.74. In behavioral research, effect sizes of this magnitude are generally considered moderate to large, suggesting that the programs did not merely nudge attitudes but were linked to substantial changes in reported behavior.

Equally striking were the secondary outcomes. Programs improved participants’ social skills with an effect size of 0.62 and boosted well-being with an effect size of 0.83, the largest effect in the analysis. This pattern supports a growing consensus in developmental science that effective anti-bullying work is not simply about punishing aggressors or warning victims. Instead, the most successful interventions build positive capacities—empathy, assertiveness, conflict resolution, emotional regulation—that displace aggressive dynamics before they take hold. Many of the included studies drew on cognitive behavioral therapy, empathy training, creative drama, peer support structures, and human-values education, reflecting the diverse theoretical traditions that Turkish researchers have adapted to their school system.

The authors were careful to flag an important statistical caveat: the effects varied considerably across studies, a phenomenon known as between-study heterogeneity. When heterogeneity is high, pooled estimates should be read as averages of a diverse landscape rather than as a single reliable prediction for any given school. To probe that variation, the team conducted exploratory moderator analyses, examining whether effect sizes differed systematically by characteristics such as educational level, participant type, and theoretical background. For bullying perpetration, effects were associated with educational level and participant type; for victimization, effects were associated with the theoretical background of the program. However, the researchers urged caution, noting that some subgroups contained few studies and that moderators were examined in separate models rather than simultaneously, which limits how confidently these patterns can be generalized.

One of the more practically useful findings emerged from exploratory univariable meta-regression analyses: session duration was associated with bullying effect sizes, whereas the number of sessions was not. In other words, what seemed to matter was how long students spent in each intervention session, not simply how many sessions were scheduled. This distinction carries real implications for program design. A school might assume that spreading a curriculum across many brief meetings is sufficient, but the data hint that sustained, in-depth engagement within each session may be the active ingredient. The finding aligns with mechanistic accounts of behavior change, which suggest that skills such as empathy and self-regulation require enough practice time within a session to consolidate.

The Turkish context makes the study particularly valuable for the international literature. Global meta-analyses, including influential syntheses by Farrington and Ttofi and updated reviews by Gaffney and colleagues, have established that school-based anti-bullying programs work on average, but questions persist about whether findings from Finland’s KiVa program, Australia’s Friendly Schools initiative, or Italy’s No Trap! model transfer to other educational cultures. Türkiye presents a distinctive case: a large, centralized education system overseen by the Ministry of National Education, rising policy attention to peer bullying, and a research literature that has expanded rapidly over the past decade. By quantifying effects specifically within this system, the new analysis provides evidence that anti-bullying strategies can succeed outside the Western contexts where most evaluations have historically been conducted.

The scope of the included studies is itself notable. Participants ranged from preschool children as young as four to adolescents of seventeen, and the interventions spanned traditional bullying and cyberbullying alike. Some programs targeted gifted students, children with specific learning difficulties, and refugee children under temporary protection, underscoring that bullying risk is not evenly distributed and that tailored approaches may be needed for vulnerable groups. Others tested delivery innovations, from drama-based sensitivity training to web-based psychoeducation grounded in cognitive behavioral principles. This breadth allowed the meta-analysis to speak to a question that single trials cannot: whether the anti-bullying enterprise, taken as a whole, is producing consistent benefits across ages, formats, and populations within one national setting.

The authors emphasize that their results underscore the importance of developing comprehensive, theoretically informed, and culturally responsive anti-bullying strategies. The moderator findings about theoretical background suggest that the conceptual engine of a program—whether it targets cognitions, emotions, values, or peer group dynamics—may shape how well it reduces victimization, even if the analysis cannot yet pinpoint which theory wins. The researchers also frame their work as informing adaptation beyond Türkiye: by identifying potential mechanisms, such as extended session duration, and boundary conditions, such as educational level, the study offers a roadmap for practitioners elsewhere who must decide whether and how to import or adapt programs designed in another cultural milieu.

Limitations remain, and the authors are transparent about them. The moderator analyses were exploratory, some categories rested on a small number of studies, and the high heterogeneity means that individual programs will vary in effectiveness. The meta-analysis also relied on previously published studies, which may themselves differ in methodological rigor. Still, the headline conclusion is difficult to dismiss: across dozens of trials and more than 1,500 children and adolescents, school-based interventions in Türkiye were consistently associated with less bullying, less victimization, better social skills, and greater well-being. For policymakers weighing investment in anti-bullying infrastructure, and for researchers designing the next generation of programs, the message is that these interventions are worth the effort—provided they are grounded in theory, given enough time in each session to work, and adapted with care to the students they are meant to protect.

Subject of Research: The effectiveness of school-based anti-bullying intervention programs in Türkiye, examined through a meta-analysis of 35 studies.

Article Title: What Works Against Peer Bullying? A Meta-Analysis of School-Based Interventions in Türkiye

Article References: Haspolat, N. K., Ağırkan, M., Ciminli, A., & Koç, M. (2026). What Works Against Peer Bullying? A Meta-Analysis of School-Based Interventions in Türkiye. Educational Psychology Review, 38(1), Article 119. https://doi.org/10.1007/s10648-026-10214-8

Image Credits: AI Generated

DOI: 10.1007/s10648-026-10214-8

Keywords: peer bullying, bullying prevention, meta-analysis, school-based interventions, victimization, social skills, well-being, cyberbullying, Türkiye, educational psychology, effect size, random-effects model