How can Adolescent Aggression be Reduced? A Multi-level Meta-analysis


Laura Castillo-Eito, University of Gloucestershire, United Kingdom
Christopher J. Armitage, University of Manchester, United Kingdom
Paul Norman, University of Sheffield, United Kingdom
Marianne R. Day, University of Sheffield, United Kingdom
Onur C. Doğru, University of Sheffield, United Kingdom
Richard Rowe, University of Sheffield, United Kingdom


Aggressive behaviour among adolescents has important social and economic costs, with negative consequences for both victims and perpetrators. Numerous psychosocial interventions have been developed to prevent and reduce aggression among adolescents. Reviews of these interventions have concluded that they are generally effective, especially when they are targeted at adolescents at risk of being aggressive. However, little is known about which intervention components enhance or diminish intervention effectiveness. In this study, we aimed to identify which techniques were more effective reducing aggression for all adolescents and for adolescents at-risk. In addition, we studied which intervention, participants and study characteristics moderated the effect.


Sixteen databases were searched for randomised controlled trials that evaluated interventions addressed to adolescents with a measure of aggressive behaviour. After screening 9795 records, 95 studies were included, which produced 283 effect sizes from 115 intervention groups. The Behaviour Change Technique Taxonomy was used to identify the techniques used in each intervention. Effect sizes were extracted for each intervention group, relevant outcome and follow-up. To account for the likely correlation between effect sizes extracted from the same study, a three-level random-effects meta-analysis was performed using the Metafor package for the R environment.


Included studies comprised 111,151 young people with a mean age of 13.36 (60% male). The interventions varied in duration from 20 minutes to 3 years, 79% of them were delivered in schools and 84% were group interventions.  There was a significant overall small-to-medium effect size (d=0.28; 95% CI [0.17, 0.39]). Significant moderator effects were found for duration of the intervention, country where the trial was conducted, level of risk of the adolescents, intervention facilitator and facilitator training. More effective interventions were of shorter duration, were conducted in the Middle East, were targeted at adolescents with higher levels of risk, were facilitated by intervention professionals and the facilitators did not get any training.

The interventions included between 0 and 22 behaviour change techniques (m = 8). However, the number of techniques did not moderate the effect. The most frequently included technique was behavioural practice. Interventions addressed to all adolescents regardless of risk were more effective if they included behavioural practice or problem solving. The technique with the biggest effect size for interventions addressed at adolescents at risk was action planning. However, interventions including action planning were not significantly more effective.


Psychosocial interventions are effective in reducing adolescent aggression, especially if there are addressed at adolescents at risk of being aggressive and if they are facilitated by an intervention professional. Interventions addressed to all adolescents should include behavioural practice and problem solving to increase effectiveness. More studies are needed to investigate the minimum duration for an intervention to be effective and to identify which techniques make targeted interventions effective.

Link to paper > Contact > Website >
Back to Spotlights