Abstract
Most empirical applications of the Bergen Social Media Addiction Scale (BSMAS) rely on student or convenience samples and often treat screentime as the main behavioral proxy. This article addresses both limitations using a representative CAWI survey of 2,033 adults in Germany conducted in May 2025. Building on clinically informed BSMAS groupings, we examine the population distribution of problematic use, the suitability of a generalized conflict item for respondents outside work or study, the usefulness of a single-item measure of platform-specific addictive pull across 18 social media services, and the comparative value of engagement-based versus frequency-based indicators. Problematic use appeared to be linked to age. Only 2.2% of respondents fell into the addicted category, but 12.3% were classified as at-risk, with markedly higher shares among younger cohorts. The amended BSMAS item for respondents not working or studying showed comparable psychometric performance to the standard wording, preserving reliability, inter-item correlation, and excellent one-factor model fit. The single-item measure of addictive pull showed a moderate positive association with the global BSMAS and allows parsimonious within-subject comparisons across platforms, although it should not replace the full scale. Most importantly, posting- and reaction-related engagement were more strongly associated with BSMAS than service intensity and, unlike intensity, significantly distinguished low-risk from at-risk users. The findings suggest that problematic social media use in the adult population is neither ubiquitous nor well captured by screentime alone. For research, prevention, and platform governance, engagement dynamics and social feedback appear to provide a more informative basis for identifying users approaching higher risk than exposure-based measures alone.