Maartje Boer

SMU AND ADHD-SYMPTOMS 131 5 factor for developing addictions, such as substance dependency (Cyders & Smith, 2009; Ohlmeier et al., 2008). Because SMU problems are characterized by addiction-like behaviors, adolescents with ADHD-symptoms may also be sensitive to developing SMU problems. We therefore expected that high levels of ADHD-symptoms increase SMU problems over time (H2). The Influence of SMU Intensity and SMU Problems on ADHD-symptoms Adolescents who intensively use social media may be accustomed to task- switching between media activities and other (offline or online) activities (Karpinski et al., 2013; Rosen et al., 2013). This may impair their ability to filter relevant from irrelevant information, which may, in turn, contribute to the development of attention deficits (Baumgartner et al., 2017). Also, intensive social media users may become habituated to the entertainment provided by social media. As a result, they may perceive activities without media that require prolonged attention as unentertaining or boring, resulting in experiences of attention deficits (Nikkelen et al., 2014). Furthermore, intensive SMU may disrupt sleep due to intensive exposure to bright screens (Van der Schuur et al., 2018), which, in turn, could lead to more attention deficits or to impaired abilities to forego immediate impulses at daytime (Fallone et al., 2001). We thus expected that SMU intensity increases ADHD-symptoms over time (H3). Also, adolescents with SMU problems may experience attention deficits due to their preoccupation with social media. Their constant urge to go online may make them feel restless when they cannot immediately check and respond to incoming messages, for example, at school. We therefore expected that SMU problems increase ADHD-symptoms over time (H4). Current Study The current study investigated the directionality of associations between ADHD-symptoms and both SMU intensity and SMU problems, using three waves of longitudinal data on Dutch secondary-school adolescents aged 11-15 years (Van den Eijnden et al., 2018). To address directionality, we applied the ‘random intercept cross-lagged panel model’ (RI-CLPM) (Hamaker et al., 2015). This novel modelling technique allowed us to examine relations between