Scrolling Past the Evidence: A Methodological Critique of Research on Social Media and Adolescent Mental Health

Public discourse surrounding adolescent social media use has increasingly positioned digital technology as the primary driver of deteriorating youth mental health. However, the academic literature does not entirely support this conclusion. This paper reviews evidence from systematic reviews, meta-analyses, large cohort studies, and ecological momentary assessment research to argue that the apparent association between social media use and adolescent mental health outcomes is consistently weak in magnitude and heavily dependent on how exposure is measured. Broad correlational studies identify negative associations in the range of r = -.05 to -.10, effects too small to be perceptible at the individual level. These modest associations are further inflated by well-documented methodological problems such as inaccurate self-reported screen time, conceptually blunt aggregation of qualitatively distinct online activities, demand characteristics in single-respondent survey designs, and analytic flexibility across model specifications. When stronger designs are applied (within-person multilevel models and time-use diary measures), associations between daily technology use and mental health symptoms largely disappear. Content-specific research further suggests that the type of social media activity, rather than total time spent, is the more meaningful predictor of outcomes. Taken together, this evidence calls for greater methodological rigour in future research and caution in translating weak correlational findings into broad policy recommendations.