Clinical Trial of Second StepŠ Middle-School Program: Impact on Bullying, Cyberbullying, & Sexual Harassment Perpetration

Results of a three-year cluster-randomized clinical trial of Second Step: Student Success Through Prevention (SS-SSTP) Middle School Program (Committee for Children, 2008) on reducing bullying, cyberbullying, homophobic, and sexual harassment perpetration. Thirty-six schools in Illinois and Kansas were randomly assigned to the intervention or control conditions. Teachers implemented 41 lessons (6^th - 8^th grade curriculum) that focused on social emotional learning skills, including empathy, bully prevention, and problem-solving skills. All 6^th graders (n = 3658) in both conditions completed self-report measures assessing verbal/relational bullying, cyberbullying, homophobic name-calling and sexual violence victimization/perpetration at four waves.  At Level 1 (i.e., time), we captured the intercept (i.e., baseline rate) and slope (i.e., change over time) in delinquency across waves 1-3. At Level 2 (i.e., student), we used the delinquency intercept (not pictured in Figure 1) and slope to predict change in bullying and other aggressive behaviors (i.e., cyber bullying, sexual violence, and homophobic teasing) between waves 1 and 4. These outcome variables were regressed on baseline levels to ensure that we were capturing the change between waves 1 and 4. At Level 3 (i.e., school), we captured the impact of the Second Step program (and school-level covariates) on the delinquency intercept and slope. To avoid underspecifying the model, each outcome variable at level 2 (i.e., bullying cyberbullying, sexual violence, and homophobic teasing) was regressed on all four baseline measures as well as both the intercept and slope terms from the delinquency trajectory. Within each level, predictors were allowed to correlate with one another, as were outcomes.  Our direct effects analysis yielded no significant findings.  Turning to our full model, we find that Second Step had a significant effect on trajectories of delinquency, which in turn were significantly related to change in bullying and related aggressive behaviors from waves 1-4. Baseline (intercept) levels were also uniformly related to each aggressive behavior, as was homophobic teasing; thus, delinquent behavior and homophobic teasing may be viewed as “early indicators” of risk for later bullying and aggressive behavior. The indirect effects of Second Step on bullying and aggressive behavior were statistically significant.  Thus, these findings suggest that Second Step reduced bullying and other forms of aggression perpetration through a reduction of delinquency.