Maartje Boer

CHAPTER 6 160 direction; neither directly nor indirectly. Rather, we expected that addiction- like SMU problems would decrease mental health. We also expected, reversely, that poor mental health would increase SMU problems. We also examined whether these proposed bidirectional associations were mediated by upward social comparisons, cybervictimization, decreased face-to-face contact with peers, and worsened school achievements (Figure 6.1). Methods Sample Data were obtained from the Digital Youth-project: a self-report longitudinal study on online behaviors and mental health among Dutch secondary school students (Van den Eijnden et al., 2018). We used data from the second, third, and fourth wave, which took place in February and March of 2016, 2017, and 2018, respectively. Data from the first wave were excluded because depressive symptoms were not measured in this wave. The waves that were included in the current study are further referred to as T1, T2, and T3. In order to study developments of adolescents from a similar age category, we selected students who were in the first two school years of secondary school at T1 ( n = 2,228). Students for whom data were missing on all study measures were excluded from the sample, which yielded a final analysis sample of 2,109 adolescents from 9 schools. From this sample, 77.9% participated in T1, 75.0% participated in T2, and 40.5% participated in T3. The nonresponse was mainly due to dropout of schools and classes, because teachers were absent or not able to schedule the survey assessments at participating schools. Hence, we considered the dropout as not selective. At T1, participating students from the analysis sample were between 10 and 16 years old ( M = 13.1, SD = 0.8) and 43.1% were first year students. In addition, 43.1% were girls, 25.7% had an immigrant background, and students were attending education at different levels (65.3% pre-vocational, 24.2% intermediate, and 10.5% pre-university). Girls and students with pre-university education were somewhat underrepresented compared to the Dutch adolescent population of the same age category in 2017 (49.1% girl, 51.1% pre- vocational, 22.1% intermediate, and 21.6% pre-university) (Central Bureau for Statistics, 2019a). Sample characteristics in T2 were approximately the same