Predictors of bullying victimization at school: a network approach

Abstract

Although school bullying is a prevalent worldwide phenomenon with enduring negative consequences on individuals’ health, social interactions, and economic well-being, bullying detection and mitigation have been challenging. Using a large dataset on high-school students (n = 3, 827) that includes information on bullying victimization and their friendship and enmity networks, we show that both networks are strongly predictive of victimization. Although a number of network features predict victimization, the primary determinants are the number of friends and enemies. Friendship ties protect against bullying while enmities expose to victimization, but the marginal effects of both variables diminish with the number of friends and enemies. Since we further document that the role of networks is largely independent of conventional correlates of victimization, our findings underscore the unique role played by socialization patterns in predicting instances of bullying among children and adolescents, offering an effective and non-intrusive method for detection and prevention efforts.