Two-Step Mediation of the Impact of a Social Emotional and Character Development Program On Behavioral and Mental Health: Results From a Randomized Controlled Trial

Introduction: The Positive Action (PA) program is a comprehensive, school-based, social-emotional and character development program with demonstrated effects on behavioral and mental health outcomes.  In the Hawaii randomized controlled trial (RCT) of PA, the program’s impact on substance use, violence, and sexual activity were mediated by positive academic behavior (Snyder et al., in press). Using data from the first RCT of PA in a low-income, urban setting (Chicago), we expand on the Hawaii analyses by testing a two-step mediation model to determine whether PA’s impact on behavioral and mental health outcomes was mediated by changes in attachment to teachers and disaffection with learning.

Methods: The fall 2004 through spring 2010 PA Chicago RCT was longitudinal at the school level and used a place-focused intent-to-treat design with a dynamic cohort (Vuchinich et al., 2012) of students followed from grades 3 to 8. Fourteen public schools were randomly assigned from matched pairs to PA or wait-listed control. For the 1,170 participating students, we tested a two-step mediation model to examine whether changes in attachment to teachers and endpoint disaffection with learning mediated PA's impact on endpoint substance use, violence, bullying, disruptive behaviors, anxiety, and depression. To address issues of non-normality, these analyses were completed using bootstrapping.

 Results: PA was associated with a positive change in attachment to teachers, and changes in attachment to teachers were inversely associated with disaffection with learning, which was positively associated with negative outcomes. The indirect effect of PA through the slope of attachment to teachers and disaffection with learning was significant for bullying (p = 0.046) and disruptive behaviors (p = 0.046), and marginal for depression (p = 0.099) and substance use (p = 0.067). These models each demonstrated strong model fit (i.e. RMSEA less than 0.05)

Conclusions: These results expand upon prior work by Snyder and colleagues (in press) by adding another, theoretically justified, mediator into the explanatory pathway.  Understanding the mechanisms by which PA exerts its influence is of interest to a multitude of stakeholders (e.g., school officials, policy makers, researchers, etc.).  Implications will be discussed.