Esther Mertens

156 | Chapter 7 only puts effort in evaluating those interventions that have a solid theoretical basis regarding the working mechanisms. In contrast, the upcoming research approach (examining intervention components) treats interventions as packages build from components that can be studied in isolation, with the main focus on “how does it work” (Collins et al., 2007). Potentially meaningful components are identified based on developmental theories or mechanisms of change. Subsequently, the effectiveness of these components are examined in an experimental design, for instance by means of a factorial experiment (Collins, Dziak, Kugler, & Trail, 2014). The focus on theory and working mechanisms in such a research approach can facilitate generalization and adaptation of intervention to other contexts. A crucial requirement to enable generalization and adaptation is an accurate and detailed description of components. Future research can only build on previous research when components are described using consistent terminology and enough information (Michie et al., 2009). When components are identified as effective, they can be used to alter existing interventions or design new interventions. However, these altered or new interventions are not necessarily improved or effective. Combining intervention components might influence their effectiveness, for instance due to interplay among components or the sequence of the individual components (Collins et al., 2007). Therefore, it is essential that research also evaluates the intervention effects of these altered or newly designed interventions in a real- world setting. I argue that intervention research can benefit from combining the more traditional research approach with the component research approach. The RCT conducted for this dissertation showed a conservative example of how the two approaches can be combined. By evaluating the intervention implemented with various levels of ecological focus, I indirectly examined the two structural components of a whole school approach and parental involvement. However, I would recommend future research to link the two approaches more explicitly and rigorously, for instance using the design proposed by Collins and colleagues (2007) consisting of three stages. First, the screening phase, relevant intervention components are identified and their effectiveness tested. Second, the refining phase, possible interactions between components and characteristics that might affect effectiveness are explored. Third, confirming phase, an intervention is build based on the evidence-base for components acquired in the previous two stages and evaluated. Hence, combining the two research approaches provides the opportunity to build interventions based on solid knowledge about working mechanisms (i.e., components) allowing for dissemination to different contexts while maintaining the necessary characteristics of the intervention (Bonell et al., 2012; Michie et al., 2009). Schools and professionals. With the current dissertation I hope to contribute to increasing awareness of school staff and professionals that intervention effects might differ between contexts (e.g., intervention dosage), participants, in particular students (e.g., personality traits), and developmental domains (e.g., intrapersonal vs.