Bullying research has grown tremendously throughout the years, and yet, there is a lack of research investigating the biological underpinnings of bullying victimization and perpetration. The single nucleotide polymorphism catechol-O-methyltransferase (COMT) Val158Met is an important candidate gene that has been demonstrated to interact with environmental factors and play an important role in emotion processing. However, it remains unknown whether COMT Val158Met influences youth and adolescents’ involvement as both targets and perpetrators of bullying, considering bully-victims are found to struggle with emotion regulation. To address this knowledge gap, the role of COMT Val158Met on the joint trajectories of bullying victimization and perpetration was investigated in a longitudinal community sample. A latent class growth analysis (LCGA) was used to identify distinct patterns of bullying victimization and perpetration across the ages 10 to 18 years (n = 648). A three-class solution was chosen for bullying victimization where most participants were reflected in a trajectory of low decreasing bullying victimization (74%), followed by moderate stable trajectory of bullying victimization (23%) and a final group following a high stable trajectory of bullying victimization (3%). A two-class solution was chosen for bullying perpetration. As predicted, most participants were reflected in the low stable bullying perpetration group (83%) and a small group followed a moderate increasing/decreasing trajectory of bullying perpetration (16.4%). Dual trajectory models revealed distinct subgroup of individuals involved in bullying either as targets, perpetrators, or bully-victims. Conditional probabilities results suggest that highly victimized youth would in time perpetrate against others while remaining targets of high levels of perpetration (i.e., target to bully-victim), whereas youth moderately victimized were more likely to be uninvolved in bullying perpetration. There was no significant difference in allelic variations (i.e., any Met allele vs Val/Val) of COMT Val158Met between bully-victims and children uninvolved in bullying. Implications of these findings are discussed from a differential susceptibility model. Gaining an understanding of the mechanisms behind the impact of bullying victimization and perpetration on children and adolescent will help provide insight and support for school and clinical prevention and intervention efforts.

Bullying is a universal problem affecting the emotional, social, and physical wellbeing of school-age children worldwide. Individual level correlates of bullying have been well-documented; however, there is limited research identifying variables at the school level which contribute to bullying involvement, especially among high school students. In this dissertation, school characteristics associated with bullying were investigated using an ecological systems framework. In the first paper, a comprehensive review of the bullying literature was conducted. Research in the following areas were summarized: definitions of bullying, measures of bullying, individual correlates, influences of cognitive development and social context across age groups, contextual variables (family, school, and community), evidence-based interventions, and bullying from a socio-ecological perspective. In the second paper, research findings are presented for an original study investigating school level predictors of bullying involvement across Kentucky high schools. The study used aggregated data from a survey of 9th to 12th grade students in 26 high schools across the state, combined with existing school datasets, in order to examine: (1) the prevalence of bullies, victims, and bully-victims across Kentucky high schools and (2) school characteristics associated with elevated rates of bullying involvement. Results revealed important differences in school bullying incident reports and student reports of bullying experiences, as well as unique differences between school environments with high and low rates of bullying involvement. Overall, academic performance and parent involvement were the strongest predictors of bullying involvement at the school level; however, the relationships between these variables and prevalence rates were not as expected. In several analyses, individual level findings from the bullying research did not translate to the school level as hypothesized. Overall, these findings have important implications for researchers when using multilevel analyses in the school context, when investigating the impact of bullying interventions at the school level, and when investigating how the school environment contributes to bullying. Results also provide important information for schools developing or revising bullying data collection procedures.