Maartje Boer

CROSS-NATIONAL VALIDATION OF THE SMD-SCALE 65 3 Reliability Reliability was assessed based on the internal consistency of the scores on the nine items using the validation subsamples. Given the dichotomous nature of the nine items, we computed the internal consistency using the tetrachoric correlation matrix, referred to as the ordinal alpha (Gadermann et al., 2012). An alpha of 0.80 or higher indicates good reliability (Gadermann et al., 2012). Measurement Invariance Measurement invariance means that the scale measures the same underlying construct across subpopulations, which is required in order to reliably compare the level of problematic SMU across subpopulations (Van de Schoot et al., 2012). To do so, we examined whether the factor structure was comparable across countries (44 countries), gender (boy and girl), age groups (11-, 13-, and 15-year-olds), and socioeconomic status (low, middle, and high family affluence) using multigroup CFA. We compared the model fit of a multigroup CFA where all item factor loadings and thresholds were free to vary across countries or subgroups (i.e., configural invariance ) with the model fit of a multigroup CFA where all item factor loadings and thresholds were constrained to be equal across all countries or subgroups (i.e., scalar invariance ) using the default model settings (L. K. Muthén & Muthén, 2017c). A test of loading invariance where thresholds are freely estimated (i.e., metric invariance ) was not conducted because this model is not identified when using dichotomous items (L. K. Muthén & Muthén, 2017c). Measurement invariance was established when the scalar model decreased CFI by not more than 0.010 and increased RMSEA by not more than 0.015, relative to the configural model (F. F. Chen, 2007; Cheung & Rensvold, 2002). Criterion Validity Criterion validity refers to the extent to which a construct relates to another construct that it should theoretically be related to. Research suggests that problematic SMU impairs mental health (Boer, Stevens, Finkenauer, De Looze, et al., 2021; Raudsepp, 2019), and that problematic users also use online communication intensively (Frost & Rickwood, 2017; Marino et al., 2018a). Accordingly, review studies show a small to moderate negative association