Bully-Victims: Evidence of a Vicious Cycle in School Bullying Among Korean Male Youth through Autoregressive Cross-Lagged Modeling

Background: According to routine activity theory, youth who engage in antisocial behavior are more vulnerable to becoming future victims, as they tend to be in high-risk situations that expose them to other potential offenders. Conversely, social learning theory suggests that victimized youth may learn antisocial behavior from bullies and end up perpetrating other youth. Youth who experience both processes—“bully-victims”—are often smallest in size, but are at greatest risk in terms of adjustment. To design appropriate intervention programs which consider the developmental and dynamic aspects of school bullying, it is fundamental to identify not a tendency, but a causal effect, in the relationship between bullying and victimization based on longitudinal data and advanced methodological approaches. The current study hypothesized that bullying and victimization behaviors are multidirectional: There are cross-lagged effects that create youth who are both perpetrators and victims of school bullying.

Methods: Analysis were based on waves 3-5 of the Korea Children and Youth Panel Survey, a nationally representative data of primary and secondary school aged male students in South Korea (N=1,167). The study examined males only due to their higher likelihood of becoming bully-victims than females. We employed autoregressive cross-lagged models to identify the causal association between bullying (e.g., experience of threatening, beating, robbing, and mocking) and victimization (e.g., experience of being threatened, beaten, robbed, and mocked) in longitudinal data.

Results: As hypothesized, there were reciprocal relationships between experience of bullying and victimization. The model fit indices reported indicate a good fit of the final model to the data ( = 14.01, df= 8, p<.08, CFI=0.96, TLI=0.93, RMSEA=0.03). Cross-lagged effects of current bullying on future victimization were statistically significant ( =.06, p<.05 for Bul_T1→Vic_T2; =.19; p<.05 for Bul_T2→Vic_T3). Similarly, cross-lagged effects of current victimization on future bullying were significant ( =.06, p<.05 for Vic_T1→Bul_T2; =.20, p<.05 for Vic_T2→Bul_T3). Lagged effects are also statistically significant between each time points such that current bullying caused future bullying ( =.10, p<.05 for Bul_T1→Bul_T2; =.38, p<.05 for Bul_T2→Bul_T3) current victimization led to future victimization ( =.24, p<.05 for Vic_T1→Vic_T2; =.64, p<.05 for Vic_T2→Vic_T3). 

Conclusions and Implications: This study used autoregressive cross-lagged modeling to identify the causal relationship between bullying and victimization. Guided by routine activity theory and social learning theory, we identified significant cross-lagged effects between bullying and victimization among Korean male youth: Current bullies showed higher levels of being victimized at later time points and current victims reported higher levels of engaging in antisocial behavior toward their peers at later time points. The risky environment in which bullies are often situated may expose them to other potential offenders and make youth vulnerable to future victimization. Concurrently, as victims interact with perpetrating youth, they may develop positive attitudes toward violence and learn violent behavior. Findings from this study may have implications for designing policies against school bulling. Not only is short-term intervention for handling immediate psychosocial maladjustment important, but so are long-term plans that prevent youth from falling into the vicious cycle of perpetration and victimization in the system of school bullying.