Esther Mertens

20 | Chapter 1 influences on the collective classroom level (Busching & Krahé, 2020). Given that I was interested in peer influences at group level (i.e., in the classroom), I analyzed the data with multilevel mediation models in which effects within individuals can be separated from effects between clusters (i.e., classrooms; Preacher, Zyphur, & Zhang, 2010). The second way I studied working mechanisms is through intervention components. Components of interventions are aspects associated with the content, instructional method, or structural characteristics of the intervention (Boustani et al., 2015; Lee et al., 2014) that affect participants’ change. I examined all three types of components: Content, instructional, and structural components. Content components represent specific skills that are taught, such as emotion regulation (e.g., Boustani et al., 2015). Instructional components represent methods used for information delivery, such as practice during the lessons (e.g., Boustani et al., 2015). Structural components represent the structure of interventions, such as the number of sessions (e.g., Lee et al., 2014). Knowing which components are necessary to establish positive intervention effects could enhance more efficient and cost effective implementation of interventions by implementing effective and eliminating ineffective components (Michie et al., 2009). Optimization of interventions is particularly important for universal school-based interventions aiming to stimulate competencies and prevent problems in the intra- and interpersonal domain, since these interventions tend to show only small positive effects (Durlak et al., 2011). Additionally, schools have limited time and resources to invest in interventions, emphasizing the need to implement optimized interventions. Therefore, I conducted a meta-analysis to identify which components were related to stronger, or weaker, intervention effects of universal school-based interventions fostering students’ development in the intra- and interpersonal domain. The results of this meta-analysis may help schools to make more informed decisions about which intervention to implement and can catalyze hypotheses generation in research regarding potentially interesting components (Wilson & Lipsey, 2007). Outline of the Present Dissertation To fill the gaps in the current literature, the present dissertation aims to gain insights into the effectiveness of school-based interventions stimulating competencies and preventing the development of problems in the intrapersonal and interpersonal domain. To this end, I examined 1) the effectiveness of these universal school-based interventions (i.e., what works?), 2) whether heterogeneity in the contexts and in the student population affected intervention effects (i.e., under what circumstances and for whom does it work?), and 3) working mechanisms of universal school-based interventions (i.e., how does it work?). Regarding the first aim, the study protocol of the RCT to evaluate R&W is described in detail in Chapter 2. This chapter uses the initial concepts “socio-emotional