Stability and constancy of bully-victim behavior: Looking at variables and individuals

The present paper (1) focused on variables to investigate the stability of bully-victim behavior over time and constancy across settings in preadolescent and adolescent boys and girls and (2) looked at persons to examine whether patterns of change were the same or different between bully-victim groups identified. Data were drawn from two independent short time longitudinal studies. In Study 1, 100 adolescents (59% female) aged 15-19 years (M = 16.4 years) were asked about their bully-victim behavior in school before and after the summer break. In Study 2, 116 preadolescents (49% female) aged 9-15 years (M = 12.2) were asked about their bully-victim behavior in school and in a summer camp. While adolescents did not enter a new group after the summer break, preadolescents entered a new group in the summer camp. On the whole sample level, both over time and across settings bullying showed moderate to high stability and a decrease. Victimization was moderately stable over time but rather unstable across settings. Victimization also decreased over time. With longitudinal cluster analysis (L-CA) both over time and across settings four bully-victim groups were identified: uninvolved, desisting victims, desisting moderate bullies, and escalating high level bullies. In both studies, escalating high level bullies were only boys. Bullying might be used to establish dominance by only a small group of boys when entering a new group or when returning to an old one after some break. Application of the variable-oriented approach using correlation coefficients and MANOVAs and the person-centered approach using L-CA and Configural Frequency Analysis complements each other.