School and classroom effects on bullying and peer victimization.

Rationale. Most studies about bullying focused on individual characteristics of bullies and victims. Only very few studies have investigated the effect of school and classroom factors on bullying. These studies indicated that between-classroom variance is higher than between-school variance. From theoretical and practical points of view, one key issue is to know if those school and classroom effects are related to educational practices rather than to the composition of the student body. At the school level, available studies only controlled for school size and mean socio-economic status. At the classroom level, composition effects have not been tested. Regarding educational practices, existing studies relied on different theoretical background and measured different dimensions: bullying and peer victimization were found to be associated with performance goal structure, teacher-student relationships, autonomy-support, structure and support. In the absence of integrative studies, it is hard to identify from those findings which kind of educational practices is more directly related to bullying and victimization. Aims. The aims of this study were (a) to assess the size of school and classroom effects on bullying/victimization, (b) to investigate the contribution of school and classroom composition to these effects, and (c) to test the contribution of a variety of educational practices to between-schools and -classrooms differences. Method. Three thousands two hundreds and sixty-three students in grade 9 (mean age 15) from 234 classrooms nested in 64 urban schools completed an anonymous questionnaire about socio-demographic characteristics, teachers practices, and involvement in bullying and peer victimization. Results. Age, grades, parental education level, parental employment, family structure, family wealth, ethnicity, and language spoken at home, were not correlated with bullying and victimization. Male reported more bullying than female. Multilevel analyses showed no significant school effects and small classroom effects on bullying and victimization. At the classroom level, composition effects were null to very small. Classroom management and competitive goal structure explained between-classroom variations in bullying and victimization. These results suggest that interventions targeted at daily classroom practices could be a way to prevent or reduce school bullying and peer victimization