Maartje Boer

CHAPTER 3 76 problematic users may engage in a high intensity of online communication, there may also problematic users who do not show intensive online communication. These latter users may experience a mismatch between their desired and actual online social network: they could be preoccupiedwith social media without having the desired network to interact with. Conversely, adolescents engaging in intensive online communicationmay be well able to regulate their online activities without experiencing problematic SMU. Strengths and Limitations The present study has several strengths, related to the data that includes many nationally representative subsamples. However, there are also some limitations that should be acknowledged. First, the cross-sectional design of the study precludes the possibility to investigate the predictive validity and the test-retest reliability of the scale. Second, the present study included mainly European adolescents. Third, other elements of validity, including convergent and discriminant validity, were not assessed. Considering these three limitations, more validation research on the SMD-scale using longitudinal data and data from non-European adolescents, and including more validation analyses, is warranted to extend current knowledge on the psychometric properties of the scale. Fourth, scores on the SMD-scale are based on self-reports, which may deviate from assessments by others. As such, the reported prevalence rates of problematic SMU may be under- or overestimated. Research comparing self- report scores with scores from, for example, teachers or parents, is considered important. Fifth, the evaluation criteria for measurement invariance testingwere obtained fromWLSMV-estimation, whichmay not performas well as withMLR- estimation (Sass et al., 2014). However, with categorical items, measurement invariance analysis with MLR-estimation can only be conducted using Chi- square-difference tests, which may falsely reject measurement invariance due to its sensitivity to large sample sizes (F. F. Chen, 2007; Cheung & Rensvold, 2002). Sixth, the present study defined adolescents reporting six or more symptoms as problematic users. Although this definition was based on findings from latent class analyses (Boer, Stevens, Finkenauer, Koning, et al., 2021), research using clinical data is required to verify whether this definition adequately identifies problematic users, for example by comparing assessments of problematic SMU by a clinician with assessments using our used definition based on the SMD-