Maartje Boer

SMU AND ADHD-SYMPTOMS 139 5 addition, all cross-lagged paths, auto-regressive paths, within-wave (residual) correlations, and means were unconstrained over time. Results of the RI- CLPM were standardized (STD YX ) to facilitate interpretation of effect sizes. Figure 5.1 Two-Variable Random Intercept (RI) Cross-Lagged Panel Model Notes. Squares represent the computed factor scores (FS). Circles represent RIs and within-person (W) values of the respective factor scores. On the within-person level, cross-lagged paths are denoted by the diagonal arrows, auto-regressive paths by the horizontal arrows, and within-wave (residual) correlations by the double-ended arrows. Auto-regressive paths, cross-lagged paths, and within-wave (residual) correlations were estimated freely. On the between-person level, RIs were correlated. In the final analysis, this model was extended with social media use (SMU) intensity, impulsivity, and hyperactivity. Monte Carlo simulations in Mplus 8.1 were carried out to determine statistical power to reject the null hypothesis of no effect (L. K. Muthén & Muthén, 2002). Power analyses were carried out using 1000 simulated samples, a sample size of n = 543, and a Type I error rate of 0.05. The power analyses were based on our RI-CLPM including free estimation of all cross- lagged effects, auto-regressive effects, and all (residual) correlations. For