Associated factors and patterns of school bullying among school-aged adolescents in China: A latent class analysis

Adolescents involved in school bullying are a heterogeneous group. Yet less is known about school bullying patterns and their associated factors in social domains in China. The objective of the study was to use latent class analysis to identify patterns of school bullying involvement and to examine factors in individual, family, and school domains associated with each bullying pattern. Data were self-reported anonymously using a questionnaire completed by students from 3 middle schools and 4 high schools in a southern city in China (N =13,166). A latent class analysis was conducted to identify involvement patterns in school bullying victimization and perpetration. Multinomial logistic regressions were used to examine factors in social domains associated with each pattern. Seven patterns of school bullying involvement were identified: Low Victim (14.5%), High Victim (4.4%), Sexual and Cyber Victim (2.5%), Physical Bully-Victim (1.4%), Verbal Bully-Victim (3.5%), All-Type Bully-Victim (3.3%) and Uninvolved (70.4%). Alcohol use, high frequency of using social networking sites, and dis-trusting people were associated with all bullying patterns (p <.05). Gender, grade level of schooling, boarding at school, cigarette smoking, being a class leader, poor academic performance, parents both working in other cities, poor relationships between family members, father's/mother's alcohol use, parental neglect, unpopularity at school, lack of feeling safe, and no close friends were each associated with at least one school bullying pattern (p <.05). Different types of school bullying victimization and perpetration co-occur among students and multiple social domain factors were associated with bullying patterns. Findings from this study suggest that consideration of the heterogeneity of school bullying and its associated factors are critical to developing targeted prevention interventions for different school bullying patterns.