Maartje Boer

CHAPTER 5 132 social media behaviors and ADHD-symptoms over time, while controlling for all possible confounding stable characteristics, such as personality traits. The technique draws on a multilevel approach by disentangling within- and between-person variance, allowing for more accurate estimations of directionality (Hamaker et al., 2015). Method Sample To examine our hypotheses, we used the first three waves of the Digital Youth- project; a longitudinal study on online behaviors and mental health among secondary school students based on self-report measures (Van den Eijnden et al., 2018). The study was conducted in February and March of 2015, 2016, and 2017, respectively. In the first wave, 543 adolescents from the first and second year of two secondary schools participated in the study. Both schools were based in the Netherlands: one school was located in medium-sized city and the other was located in a large city. Participants were between 11 and 15 years old ( M age = 12.91, SD age = 0.73). Of this sample, 293 adolescents (54%) participated in all three waves, 198 (36%) in two waves, and 52 (10%) in one wave. Non- response was mainly due to dropout of entire school classes and not due to individual selection, because teachers were absent, or because teachers were not able to schedule time for the completion of the survey. During the first wave, school year and gender were evenly distributed (51% first year students, 52% girls). Adolescents attending pre-university education (48%) and adolescents with two Dutch parents (84%) were somewhat overrepresented compared to the composition of the Dutch adolescent population in the first two years of secondary school (26% and 73%, respectively) (Central Bureau for Statistics, 2018b). Survey participation occurred through digital self-completion during school hours and was voluntary and anonymous. Participants did not receive any incentives. Research assistants were present during assessments to assist when necessary. Participants were instructed that they were allowed to quit the survey at any timeduringassessment. Parents received information letters prior to survey participation, which provided them with the opportunity to refuse participation of their child. The study procedures were carried out in