Esther Mertens

| 155 General Discussion which components are indeed (in)effective, under what circumstances and for whom (Bonell et al., 2012). Subsequently, existing interventions can be improved by adding effective and eliminating ineffective components (Michie et al., 2009), or new interventions can be designed based on the evidence-base for components (e.g., Weisz et al., 2012). In addition, research should not only study how intervention components relate to immediate intervention effects, but also to long-term effects. Long-term intervention effects indicate that the intervention successfully andmeaningfully changed students’ development (Flay et al., 2005). However, many interventions do not succeed in changing one’s development which is illustrated by extinction of intervention effects over time (e.g., Dray et al., 2017). To improve interventions’ long-term effectiveness, it is pivotal to unravel which components contribute to long-term intervention effects and which components counteract these effects. Practical Implications Research. Effects of intervention are often studied either through evaluating complete intervention programs or through examining intervention components. These two approaches reflect two relatively separated research areas in which the first is a more traditional approach and the second an upcoming research approach (Collins et al., 2007). Both research approaches have their advantages and disadvantages. I recommend future research to combine these two research areas to maximize the advantages and minimize the disadvantages of the separate approaches. The traditional approach (evaluating complete intervention programs) treats each intervention as a whole programwith the main focus on “whether it works” (Collins et al., 2007). An RCT is considered the “golden standard” to evaluate interventions and seems indeed a solid design to test intervention effects. Interventions are developed, tested, adjusted and tested again in a new RCT (Collins et al., 2007). This research approach has led to high quality research examining sometimes interventions that have poor theory, have poorly understood effects, and are difficult to translate to other contexts (Bonell et al., 2012). As RCTs are time consuming and expensive, this process of intervention optimization is slow and inefficient. Research is focused on a specific intervention and the question remains to what extent the intervention effects can be generalized (Rowe & Trickett, 2018). At this moment, intervention databases contain many potentially effective interventions that still need to be evaluated by means of an RCT (e.g., 201 of the 233 interventions in the Database Effective Youth Interventions; Dutch Youth Institute (NJi), 2020). On top of this, new interventions are designed on a daily basis. As a result, multiple research teams are evaluating interventions with different levels of similarity without coordination between the teams (Bonell et al., 2012). To avoid waste of time and energy, it is therefore important that future research 7