Esther Mertens

18 | Chapter 1 systems). Even though manipulation of dosage is typically not feasible in evaluation studies (Farrell et al., 2013), the design of my RCT enabled me to manipulate dosage and randomly assign schools to these different conditions. The second dosage characteristic I examined concerned the time span of the intervention. It is reasonable to expect that the rate of change in participants during an interventionmight not be linear. Therefore, I conductedmeasurements before, during, and after the intervention in order to model trajectories of change and examine intervention effects over time, as recommended by Greenberg and Abenavoli (2017). Overall, with this design I was in the position to evaluate the effect of the dosage of the intervention to which students were exposed (i.e., both ecological focus and time span of the intervention) on intervention effects. As characteristic of participants, I focused on the extent to which students’ Big Five personality traits affected intervention effects. According to the vulnerability theory (Tackett, 2006), certain personality traits can put individuals at increased risk of developing problems in the intra- and interpersonal domain. Especially these more vulnerable students might benefit most from the intervention. As the Risk moderation hypothesis (Spoth, Shin, Guyll, Redmon, & Azevedo, 2006) states, vulnerable students have more room to improve on competencies or to decrease on problems, indicating the potential for a compensatory effect, whereas less vulnerable students might experience a ceiling or a floor effect (Nehmy & Wade, 2014). Furthermore, certain personality traits might enhance the transfer of skills learned during the intervention to one’s daily life facilitating generalization of the skills. Research showed that Big Five personality traits can indeed affect intervention effects. However, the results of these studies do not show a clear pattern indicating which personality traits are most important to take into account in interventions. For instance, Huppert and Johnson (2010) found stronger intervention effects on wellbeing for individuals with high levels of Agreeableness and Neuroticism. De Vibe and colleagues (2015) also found stronger effects on wellbeing for high levels of Neuroticism, but not for Agreeableness. Additionally, Wang and colleagues (2017) found no moderation of intervention effects on wellbeing by personality traits. The present study examined personality traits as moderators of intervention effects across a broad range of outcomes in the intra- and interpersonal domains aiming to reveal a pattern of moderation that could help clarifying the role of personality traits in interventions. Aim 3: Working Mechanisms in Interventions Knowledge about how an intervention establishes change is essential for understanding what is critical to an intervention and how interventions can be optimized (Kazdin, 2007). Studying working mechanisms can confirm theories, may lead to new insights for theory development (Michie, Fixsen, Grimshaw, & Eccles, 2009), or might inform directions for intervention optimization (Kazdin, 2007). In the present