Maartje Boer

CROSS-NATIONAL VALIDATION OF THE SMD-SCALE 73 3 countries (Georgia and the Russian Federation), there were no significant differences in the intensity of online communication. In one country, there was a small/moderate negative association between problematic SMU and the intensity of online communication (Azerbaijan: β = -0.273, p = 0.001). Estimates by country are presented in the Appendix (Tables A3.10-A3.12). Table 3.5 Summary Table Life Satisfaction, Psychosomatic Complaints, and Intensity of Online Communication, by Problematic SMU (n = 222,532 in 44 Countries) Means Effect size mean differences Mean 95% LL 95% UL Observed range mean 1 β SE p Coun- tries 2 Observed range β 3 Life satisfaction ( M = 7.73, SD = 2.03, min. = 0, max. = 10) Non- problematic 7.79 7.79 7.80 6.67 8.56 Problematic 6.96 6.92 7.00 6.13 8.30 -0.395 0.011 <0 .001 40 -0.682 -0.193 Psychosomatic complaints ( M = 2.08, SD = 0.90, min. = 1, max. = 5) Non- problematic 2.03 2.03 2.04 1.60 2.39 Problematic 2.62 2.60 2.63 2.06 3.26 0.648 0.010 <0 .001 44 0.309 0.924 Intensity of online communication ( M = 3.76, SD = 1.29, min. = 1, max. = 5) Non- problematic 3.72 3.72 3.73 2.84 4.12 Problematic 4.15 4.13 4.17 2.33 4.45 0.313 0.009 <0.001 41 0.163 0.635 Notes. SMU = social media use; LL = confidence interval lower limit; UL = confidence interval upper limit; β = STDY- standardized (i.e., B /standard deviation(Y)), controlled for gender, age, and socioeconomic status; SE = standard error; p = p-value. 1 Observed means across 44 countries. 2 Number of countries where a significant association was observed in the same direction as in the pooled sample. 3 Observed range STDY-standardized β across countries where a significant association was observed in the same direction as in the pooled sample, controlled for gender, age, and socioeconomic status. Overall, for almost all countries, the associations were significant and in the expected directions, which confirms appropriate criterion validity. To investigate the robustness of this conclusion, we repeated the analyses while defining problematic SMU as reporting at least five or seven symptoms, instead of six. Results were highly comparable, suggesting that our findings were not sensitive to our operationalization of problematic SMU. A summary of this analysis is provided in the Appendix (Table A3.13). Discussion The present study is the first to systematically analyze the psychometric properties of a problematic SMU-scale across comparable nationally representative samples of adolescents in many countries. Findings from