Esther Mertens

76 | Chapter 4 Adolescents and parents gave active informed consent for completing the questionnaires. In addition, parents gave passive informed consent for the participation of their child. This trial was approved by the Ethical Committee of the Faculty of Social and Behavioral Sciences of Utrecht University (FETC17-05) and registered in the Dutch Trial Register, number NL6371 (NTR6554; see for protocol Mertens, Deković, Van Londen, & Reitz, 2018). Participants At baseline, 1299 adolescents in the 7 th Grade participated in this study with an average age of 12.38 years ( SD = .62). In total, 661 (54%) adolescents were boys, and 815 (69%) adolescents had a Western background. Regarding the parents, 461 parents completed the questionnaires. Overall, parents’ ages ranged between 28 and 77 years old ( M = 43.16, SD = 5.69) and they were mostly mothers ( n = 382, 83%). The demographics are presented per condition in Table 1. Conditions did not differ on adolescents’ and parents’ sex distribution, and parents’ age. There were small differences between conditions on adolescents’ age and ethnic background (see Table 1). In the Light condition, adolescents were slightly younger than in the Control condition. Regarding adolescents’ ethnic background, in the Light and Plus conditions, the majority were adolescents with a Western background, whereas in the Standard and Control conditions ethnic background was roughly equally distributed. Missing data. Data were not missing completely at random according to the Little’s MCAR test (χ 2 (1507) = 1606.32, p = .037). However, attrition analyses showed no differences between drop-outs and completers on the demographic variables (Age: F (1, 1230) = .24, p = .626, η 2 partial  = .000; Sex: χ 2 (1) = 1.36, p = .244, φ = -.033; Ethnic background: χ 2 (1) = 1.19, p = .276, φ = .032) nor on the outcome variables, analyzed with MANOVAs per time point (T1: F (27, 3609) = 1.32, p = .126, η 2 partial  = .010; T2: F (27, 3378) = .65, p = .916, η 2 partial  = .005; T3: F (18, 2044) = .82, p = .674, η 2 partial  = .007). As we found no differences between drop-outs and completers and because the Little’s MCAR test is conservative when used on a large set of variables, we regarded the missing data as missing at random (Van Ness, Murphy, Araujo, Pisani, & Allore, 2007).