Does any of this work?
A fair question to ask an organization like ours. Here is what the research actually shows, including the numbers most organizations in this field do not put on their websites.
The short answer.
Yes, and by less than anyone selling a program will tell you.
The most comprehensive review to date pooled 100 evaluations of school-based anti-bullying programs. On average, bullying perpetration fell by 19 to 20 percent and victimization by 15 to 16 percent, with substantial variation between countries and programs.1 An earlier meta-analysis put the average reduction at about 20 percent and noted that a substantial share of programs showed no benefit at all.2
So a good program moves the number by roughly a fifth. It does not end bullying in a school, and any organization telling you otherwise is selling something.
Photo placeholder
Photo placeholderThe best-evidenced program in the world.
KiVa was developed by Christina Salmivalli and colleagues at the University of Turku in Finland. Its central idea came out of her earlier research identifying the roles children occupy around bullying: aggressors, targets, assistants, reinforcers, outsiders, and defenders.3 Rather than focusing on the child doing the bullying, KiVa works on the onlookers, on the reasoning that bullying loses its payoff when the audience stops supplying one.
The Finnish evaluation covered 8,237 students in grades four to six across 78 schools randomly assigned to intervention or control. After nine months it showed beneficial effects on seven of 11 measured outcomes, including both self-reported and peer-reported victimization.4 KiVa now runs in more than 90 percent of Finnish public schools.5
What a real-world trial found.
A large United Kingdom trial gives the most honest picture available, because it reports what did not work alongside what did.
The scale
118 primary schools across four sites, 11,111 students in the analysis.6
The effect
Victimization fell from 20.3 to 17.7 percent in program schools. Comparison schools went from 21.6 to 20.7 percent.6
What did not move
No measurable difference in bullying perpetration, school wellbeing, or emotional and behavioral problems.6
The cost
About 21 pounds more per pupil in year one, and under two pounds per pupil after that.6
Why we publish the unflattering number.
That trial found no effect on perpetration. Students did report higher empathy and fewer peer problems, and victimization did fall by a statistically significant margin. But the children doing the bullying did not measurably do less of it.
Most organizations in this field would leave that out. We put it on the page for two reasons. A parent deciding whether to push their school toward a program deserves to know what it will and will not fix. And an organization that only publishes the flattering half of a study should not be trusted with the other half.
It also points at something useful. Reducing what a targeted child experiences and changing the behavior of the child doing it are different problems, and a program that solves the first may not touch the second.
Photo placeholderWhat the evidence points toward.
One
Whole school beats one assembly.
Universal programs that change how an entire school operates outperform targeted or one-off approaches. A speaker for an hour is not an intervention.
Two
Work on the bystanders.
The most evidenced program in the world is built on this. Most children present at an incident are neither bullying nor being bullied, and what they do determines whether it pays off.
Three
Sustained, not seasonal.
The Finnish results held year on year at national scale. Programs that run for a month in October do not have the same evidence behind them.
Four
Fidelity matters.
Trials measure whether a program was actually delivered as designed, because the ones delivered loosely produce weaker results. Buying a program is not the same as running it.
Five
Cheap, relative to the harm.
Under two pounds per pupil per year after setup, against outcomes that follow children into adulthood. Cost is rarely the real obstacle.
Six
Expect a fifth, not a cure.
Plan around a meaningful reduction rather than elimination, and judge a program against that rather than against a promise nobody has met.
Photo placeholderWhere that leaves our own programs.
We should be as honest about ourselves as about anyone else. Our programs have not been through a randomized controlled trial. Very few programs outside the handful named above have.
What we can say is which evidence we built on. Facility Training follows the whole-setting finding, training every adult in a building rather than addressing children alone. Our guidance on bystanders follows the research on what onlookers actually do. The Coloring Book Campaign follows the early childhood finding that roles at that age are not yet fixed.
That is a reasoned basis, not proof, and we would rather say so than imply more.
Sources.
1. Gaffney, H., Ttofi, M.M., and Farrington, D.P. (2019). Evaluating the effectiveness of school-bullying prevention programs: an updated meta-analytical review. Aggression and Violent Behavior, 45, 111 to 133. Pooled 100 evaluations of whole-school and targeted programs.
2. Ttofi, M.M., and Farrington, D.P. (2011). Meta-analysis of manual-based bullying prevention programs, reporting an average reduction in bullying prevalence of about 20 percent, and noting that a substantial share of programs showed no beneficial effect.
3. Salmivalli, C., Lagerspetz, K., Bjorkqvist, K., Osterman, K., and Kaukiainen, A. (1996). Identification of six participant roles in bullying. See also Salmivalli, C. (2010), Bullying and the peer group: a review, Aggression and Violent Behavior, 15, 112 to 120.
4. Karna, A., Voeten, M., Little, T.D., Poskiparta, E., Kaljonen, A., and Salmivalli, C. (2011). A large-scale evaluation of the KiVa antibullying program: Grades 4 to 6. Child Development, 82(1), 311 to 330. 8,237 students, 78 schools randomly assigned.
5. Reported scale-up of KiVa across Finnish public schools following the 2007 to 2009 national trial.
6. Multicentre cluster randomized controlled trial of KiVa in United Kingdom primary schools. 118 schools, 11,111 students. Victimization 20.3 to 17.7 percent in intervention schools against 21.6 to 20.7 percent in comparison schools, odds ratio 0.87, 95 percent confidence interval 0.78 to 0.97, p equals 0.009. No significant differences found in bullying perpetration, school wellbeing, or emotional and behavioral problems. Costs reported at 20.78 pounds per pupil in year one and 1.65 pounds per pupil subsequently.