Maartje Boer

CHAPTER 8 240 Methods Data We used data from the Digital Youth (DiYo) project, which is a longitudinal survey with yearly time intervals among Dutch secondary school adolescents, conducted in 2015 until 2019 (Van den Eijnden et al., 2018). The survey assessed self-report internet-related behaviors and wellbeing. For the present study, we selected four waves of data from adolescents who were in 7 th grade at time of the 2015 ( n = 1,352) or 2016 ( n = 998) survey assessments. Adolescents who had repeated a class ( n = 46) or who participated in less than two waves ( n = 885) were excluded, which yielded an analysis sample of 1,419 adolescents. The proportion of boys, pre-vocational educated adolescents, and adolescents with an immigrant background was higher among excluded adolescents than among included adolescents. However, these differences were very small (Cramer’s V < 0.109). In addition, excluded adolescents reported lower life satisfaction at T1 until T3, higher SNS posting intensity at T1 and T2, higher SNS and IM viewing intensity at T1, lower levels of upward social comparison at T2, and more SMU problems at T1 until T3, as compared to included adolescents. Again, these differences were small (Cohen’s D range = 0.151 to 0.368). Adolescents in the analysis sample ( n = 1,419) were on average 12.51 years at T1 ( SD = 0.60), 45.95%was female, and 21.86%had an immigrant background. In the Dutch education system, adolescents are enrolled in different educational levels from 12 years onwards (i.e., when transitioning to secondary school), namely pre-vocational, intermediate, and pre-university level (57.79%, 28.54%, and 13.67%, respectively in the present study). The distributions of female adolescents and adolescents with an immigrant background in our study were approximately similar to the distributions in the 13- to 16-year- old population in the Netherlands in 2018/2019 (Central Bureau for Statistics, 2021). Adolescents enrolled in the pre-vocational educational level were slightly overrepresented (57.79% vs. 49.42%) and adolescents enrolled in the pre-university educational level were slightly underrepresented (13.67% vs. 20.62%) in our study (Central Bureau for Statistics, 2021). In T1, 44.89% of the analysis sample did not participate. In T2, this was 6.48%, in T3 24.10% and in T4 65.12%. Dropout in T1 was due to the fact that adolescents who entered the study after T1 were also included in the sample.