Evaluation of the Olweus Bullying Prevention Program: A large scale study of U.S. students in grades 3–11

Effects of antibullying programs on teachers’ interventions in bullying situations. A meta‐analysis

“…Removing the study of Howard et al () and Newman‐Carlson and Horne () results in nonsignificant summary effect sizes. The impact of antibullying programs on teachers became negligible when the study of Limber et al () was removed from the data set, indicating that our findings related to teacher interventions seems to be influenced by the outcomes of this large‐scale study.…”

“…The included studies ( n = 13) examined a total of eight antibullying interventions: KiVa (originated in Finland), Bully Busters (UK), Olweus Bullying Prevention Program (OBPP; Norway), I DECIDE (UK), Steps to Respect (US), ViSC Social Competence Program (Austria), the Sheffield Project (UK), and Expect Respect (US). Almost all studies were conducted in the country where the program was developed, except three studies from the United Studies that examined effects of the OBPP (see Black & Washington, ; Limber, Olweus, Wang, Masiello, & Breivik, ; Pepler, Craig, O'Connell, Atlas, & Charach, ). Table provides an overview of the study characteristics of the included studies.…”

“… Note : Limber et al () conducted an extended‐age cohort design. Black and Washington () and Pepler et al () conducted a longitudinal design. …”

“…Applying the leaving‐one‐out analysis results in statistically significant summary effect sizes varying between g = 0.095 (leaving out Limber et al, ) and g = 0.435 (leaving out Smith et al, ). Removing the study of Howard et al () and Newman‐Carlson and Horne () results in nonsignificant summary effect sizes.…”

“…In addition to these possible moderators at outcome level, the differences in effect sizes can also be related to differences in the research design and interventions (Higgins, Thompson, Deeks, & Altman, ). For instance, the study of Limber et al () heavily influenced the summary effect size due to the exceptionally large sample. More primary studies are needed to conduct an analysis in which moderating variables can be identified with adequate statistical power.…”