Maartje Boer

CHAPTER 6 176 cybervictimization (M1c: r = 0.187, p < 0.001), and decreased subjective school achievements (M1d: r = -0.116, p = 0.009) in the same year. In addition, adolescents with increased SMU intensity in T1 or T2 reported increased face- to-face contact within the same year (M1e: r T1 = 0.195, p < 0.001 and r T2 = 0.117, p = 0.022). However, we cannot infer directionality from these correlations. Moreover, correlations with SMU intensity in T1 may be driven by SMU problems in T1, as correlations in T1 did not take into account variance due to T1-covariates (Figure 6.2). Also, we did not find evidence that SMU problems decreased mental health via upward social comparisons and cybervictimization, because increased upward social comparisons and cybervictimization in T2 did not predict decreased mental health one year later. However, adolescents’ increases in upward social comparisons in T2 were associated with increases in depressive symptoms (M1b: r = 0.204, p = 0.001) and decreases in life satisfaction (M2b: r = -0.172, p = 0.006) within the same year. Also, increases in cybervictimization in T2 co-occurred with increases in depressive symptoms within the same year (M1c: r = 0.114, p = 0.030). Yet, we cannot infer directionality from these correlations. Gender Differences Finally, we examined whether our results were robust to gender. Researchers proposed that girls have a higher tendency to ruminate about content on social media and to compare themselves with others online than boys (Nesi & Prinstein, 2015; Underwood & Ehrenreich, 2017). Consequently, girls may be moreprone toadverseeffectsof SMU(problems) thanboys. Genderdifferences were investigated using multiple group RI-CLPMs (Mulder & Hamaker, 2021). More specifically, we estimated RI-CLPMs where all parameters were free to vary across boys and girls. In these models, we obtained z -scores for the differences in the strength of the cross-lagged parameters between the two groups. Findings of these results may be consulted in the Appendix (Tables A6.5 and A6.6). The analyses showed that although the strength and significance of the cross-lagged parameters differed slightly between boys and girls, these differences were not significant in any of the waves. These findings suggest that the associations are equally strong for boys and girls.