Maartje Boer

THE COMPLEX ASSOCIATION BETWEEN SMU AND WELLBEING 251 8 Table 8.5 Summary Results of Random Effects Models, Within-Person Associations (Level 1), Life Satisfaction Model estimates Model comparison 1 Not controlled for SMU problems Model B SE ( B ) p Var. SE (Var.) Δ Par. Δ Deviance Δ Deviance p 2 Δ AIC Δ BIC LL-B UL-B LL- β UL- β SNS viewing 5a -0.002 0.012 0.865 0.008 0.004 2 -10.504** 0.003 -6.5 6.8 -0.172 0.168 -0.296 0.289 SNS posting 5b -0.012 0.008 0.164 0.006 0.003 2 -8.568** 0.009 -4.6 8.7 -0.157 0.133 -0.271 0.231 SNS liking 5c -0.004 0.007 0.602 0.005 0.003 2 -18.406*** <0.001 -14.4 -1.1 -0.145 0.137 -0.303 0.288 SNS responding 5d -0.021 0.012 0.065 0.006 0.002 2 -17.342*** <0.001 -13.3 -0.1 -0.178 0.135 -0.361 0.275 IM viewing 5e -0.011 0.011 0.324 0.012 0.004 2 -18.098*** <0.001 -14.1 -0.8 -0.226 0.203 -0.349 0.314 IM sending 5f -0.024 0.012 0.052 0.007 0.004 2 -20.926*** <0.001 -16.9 -3.6 -0.182 0.135 -0.320 0.237 Controlled for SMU problems Model B SE ( B ) p Var. SE (Var.) SE ( B ) Δ Deviance Δ Deviance p 2 Δ AIC Δ BIC LL-B UL-B LL- β UL- β SNS viewing 6a 0.012 0.013 0.344 0.006 0.004 2 -8.726** 0.008 -4.7 8.6 -0.140 0.165 -0.242 0.284 SNS posting 6b -0.008 0.009 0.396 0.005 0.004 2 -8.914** 0.007 -4.9 8.4 -0.145 0.130 -0.253 0.226 SNS liking 6c 0.004 0.007 0.613 0.005 0.002 2 -15.502*** <0.001 -11.5 1.8 -0.133 0.140 -0.280 0.295 SNS responding 6d -0.011 0.013 0.377 0.006 0.002 2 -13.024** 0.001 -9.0 4.3 -0.161 0.139 -0.328 0.283 IM viewing 6e 0.008 0.011 0.469 0.011 0.004 2 -15.822*** <0.001 -11.8 1.5 -0.198 0.215 -0.308 0.334 IM sending 6f -0.011 0.012 0.393 0.005 0.003 2 -14.822*** <0.001 -10.8 2.5 -0.153 0.132 -0.271 0.233 Notes. SNS = social network sites; IM = instant messengers; SMU = social media use; B = unstandardized coefficient; SE = standard error, p = p- value; Δ = change relative to the baseline fixed effects models M1,2a-f ; Var. = variance of the slope; Par. = Number of free parameters; Deviance = -2*loglikelihood; AIC = Akaike Information Criterion; BIC = Bayesian Information Criterion; LL = 95% prediction interval lower limit; UL = 95% prediction interval upper limit; β = STDYX-standardized coefficient. Models 5a-f were controlled by wave (level 1) and gender, educational level, and immigrant background (level 2). Models 6a-f extended model 5a-f with SMU problems as additional control variable on the first and second level. All models included a covariance between the random slope and random intercept. 1 Models 5a-f were compared to Model 1a-f, respectively; Models 6a-f were compared to Models 2a-f, respectively. 2 The p- value for the deviance was corrected to take into account the boundary of the slope variance parameter (Hox, 2010b; Stoel et al., 2006). *** p < 0.001, ** p < 0.01, * p < 0.05.