Methods: Data are from a group-randomized trial of YM that was conducted in 28 public elementary and middle schools. Fourth-grade classrooms in participating schools were randomly assigned to receive YM or routine classroom content (N=876; 52% female; 53% Latino/a). Intervention occurred in the 4th and 5th grades, followed by a one-year follow-up after the first year of middle school. Measures used in this analysis include self-reports of bullying and peer victimization from the Olweus Revised Bully/Victim Questionnaire administered at baseline and at the end of 4th, 5th, and 6th grades. We used PCM and examined rates and patterns of involvement in bullying and/or victimization over time by following a two-step approach. Latent class analysis (LCA) was first conducted at each of the four time points. Pearson's chi-square tests were then used to examine differences in the transitional patterns of bullying and victimization among subjects by condition.
Results: LCA revealed a four-class solution at each time point with classes categorized as bullies, victims, bully-victims, and uninvolved. At time 1, bully-victims, the smallest class, comprised 11% of the sample and ‘uninvolved' had the highest percentage of subjects at 39%. We examined the transition of experimental and control subjects from the bully, victim, and bully-victim classes to the uninvolved class over the four time points. Subjects in the YM group transitioned from membership in bully, victim, and bully-victim classes to the uninvolved class at significantly higher rates than their control counterparts; these effects were strongest during the first year of intervention (grade 4) and in the transition year to middle school (grade 6). The most positive impact of YM was evident among students in the victim class; YM subjects in this class transitioned to the uninvolved class at significantly higher rates than control subjects at each time interval.
Implications: Results suggest that implementing prevention programs in the latter years of elementary school may be a promising approach to altering patterns of aggression and victimization during the transition from elementary to middle school. Findings from PCM analyses extend our prior results revealing positive effects of YM on rates of victimization and demonstrate the utility of the method for understanding the effects of interventions on sub-groups of students. PCM is discussed as a useful strategy in the context of intervention research.
