Examining the Stability in Bullying and Victimization From Middle Childhood to Early Adolescence Using Latent Variable Modeling

Background and Purpose: Bullying during childhood and adolescence is often characterized as behavior involving interactions among bullies, victims, bully-victims, and uninvolved youth. Membership and rates of participation in these four groups have generally been assessed using cross-sectional data. Consequently, little is known about the transitory nature of bully and victim groups or whether these groups are stable over time. Unfortunately, most techniques used to classify youth into bully and victim groups have relied on predetermined cut-off points that pose significant analytical limitations. To that end, the present investigation uses latent class analysis (LCA) to empirically categorize youth into latent bully and victim groups at three different time points, and then assesses the transitions across groups over time.

Methods: Data from a group-randomized trial of a bullying prevention intervention in 28 public elementary schools are used in the analysis. Measures include self-reports of involvement in bullying and victimization using the Olweus Revised Bully/Victim Questionnaire completed by participants in the study's control condition (N=674; 53% female; 53% Latino) at the end of fourth, fifth, and sixth grades. LCA results at each time point were used to examine how youth transitioned across bully and victim groups over time.

Results: LCA findings revealed that a four class solution composed of distinct bully, victim, bully-victim, and uninvolved classes best fit the data in fourth grade (time 1). Three latent classes consisting of victims, bully-victims, and uninvolved youth were identified in grades 5 (time 2) and 6 (time 3). No bully-only class was identified after the spring of the fourth grade. Results also indicate that children frequently transition from one latent class to another—only seven percent of children remained stable in a bully or bully-victim group and only 18% stay uninvolved across all three waves. Thirty-five unique transition patterns over the three grades were identified.

Conclusions and Implications: Our findings suggest that bullying and victimization incidents for many youth occur as limited episodes rather than consistently over time. Thus, it may be more accurate to think of bullying and victimization as behaviors that occur in the context of child and adolescent development rather than conduct that is specific to distinct groups of students at particular points in time. These findings are somewhat contrary to previous studies assessing change in bullying and victimization. One explanation for this difference is that children may be more likely to explore new social roles during key developmental phases such as the transition to middle school. A student's need to establish social status and relational power during such transitions may in turn increase the use of aggression as a means to achieve a more prominent position among peers. Consequently, the membership structure of bully and victim groups may be far more transient than stable. Implications of these findings for preventing bullying and victimization are discussed.