Characteristics of Responders to the Youth Matters Program: A Latent Growth Curve Analysis

Introduction: Bullying and victimization among elementary and middle school students is a persistent individual and social problem. Interventions using social and emotional learning strategies have produced positive effects (Farrington & Ttofi, 2009). However, studies assessing mechanisms of change, and identifying who benefits most from bullying prevention programs are limited. Results from a cluster-randomized trial of Youth Matters  (YM), a social and emotional learning program for elementary and middle school students, are discussed below. Although prior investigations of YM revealed significantly lower victimization rates among experimental subjects when compared to controls, the relationship between individual and social characteristics of YM participants and bullying and victimization outcomes has not been extensively examined.  

Methods: Participants included 583 students (64.5% Latino; 50.3% female) assigned to the YM condition beginning in the 4th grade. Participants were subsequently surveyed in the fall and spring for two academic years during program implementation with a 12-month follow-up conducted in the spring semester of 6^th grade. Study outcomes were levels of victimization and bullying as measured by the Olweus Revised Bully-Victim Questionnaire (Olweus, 1996). Separate linear latent growth curve models were estimated in Mplus 7.0 (Muthén & Muthén, 2012). Intercept and slope growth terms were regressed onto several independent variables: gender, ethnicity (5 dummy-coded variables; reference group: Latino/a), age, and baseline ratings of anti-social attitudes, classroom friction, delinquent behavior, and school commitment.

Results:

Victimization. Higher reported levels of baseline classroom friction were indicative of higher levels of baseline victimization (p < .01). Higher baseline levels of classroom friction were also associated with more rapid declines in victimization (p < .01). Similarly, students with higher initial reports of anti-social attitudes declined in victimization at a faster rate (p = .02).

Bullying. Female students reported significantly lower rates of bullying at baseline (p < .01). Also, students who reported higher initial levels of anti-social attitudes (p < .01) and delinquency (p < .01) reported more frequent engagement in bullying behaviors at baseline. These three independent variables were also predictive of change in bullying over time; levels of bullying among girls decreased at a slower rate compared to boys (p = .02) and students with higher initial levels of anti-social attitudes (p = .04) and delinquency (p = .04) reported more rapid declines in bullying.

Conclusions: Results suggest that the beneficial effects of YM were invariant over different characteristics.  Notably, the effect of YM differed for girls and boys where boys evidenced greater declines over time.  Similarly, correlates of bullying (anti-social attitudes, delinquency) were indicative of higher initial levels of bullying but greater declines over the course of the study. Together, these results suggest that YM is particularly effective in reducing bullying among youth who reported the highest levels of self-reported bullying at baseline.