Maartje Boer

THE COURSE OF PROBLEMATIC SMU 219 7 Limitations and Future Directions Findings of the present study should be interpreted in light of several limitations. First, we investigated adolescents’ trajectories during a limited time span, that is, four years within early and middle adolescence. During this period, social media may play a larger role in adolescents’ daily lives than in other periods, because this period typically revolves around forming new friendships, exploring new perspectives, and constructing and sharing personal narratives, which can be facilitated by social media. When these developmental tasks are (partly) fulfilled andpersonal needs change, different trajectoriesmay emerge. As such, we expect that findings from the present study cannot be generalized to older adolescents. Future research comparing trajectories of problematic SMU across younger and older adolescents would improve our understanding of adolescents’ problematic SMU in the context of their developmental period. Second, the present study only assessed trajectories of Dutch adolescents. There are substantial cross- national differences in young adolescents’ level of problematic SMU and within the European region, high levels of problematic SMU are the least prevalent in the Netherlands (Boer, Van den Eijnden, et al., 2020). As such, trajectories of Dutch adolescents may deviate from trajectories of adolescents from other cultures. Third, we used self-report measures to indicate SMU frequency, which may have limited accuracy (Junco, 2013). Researchers stressed that the frequency of SMU may bedifficult to recall and toestimate (Parry et al., 2020), which is plausiblegiven that SMU typically occurs fragmented throughout the entire day. More objective assessments would be necessary to diminish the influence of recall biases, which furthermore also reduce socially desirable responding biases. Hence, to gainmore insight into the co-trajectory of adolescents’ SMU frequency and problematic SMU, replicating our study using more objective measures of SMU, such as time tracking applications, are considered promising. Fourth, we determined the time- invariant (i.e., trait-like, stable) part of adolescents’ level of subjectivewellbeing, self- control, and social competencies based on four waves of data across four years. However, a longer time frame may facilitate more accurate estimates of trait-like psychosocial factors. Therefore, to gain more robust insights into the explanatory role of the investigated psychosocial characteristics in adolescents’ trajectories of problematicSMU,moreresearchusing longitudinal dataacrossa longer timespan (e.g., from middle childhood to late adolescence) is considered important. Fifth, because participating schools were not sampled through a random sampling