Author Name: Madhav Menon Date: 27-03-2026
The relationship between social media exposure and anxiety has become one of the most contested empirical questions in contemporary psychological science, with studies producing contradictory findings that reflect methodological heterogeneity rather than genuine theoretical inconsistency. This paper provides a comprehensive theoretical and empirical review of the social media-anxiety association, with specific focus on the measurement challenges that have prevented resolution of the dose-response question: does more social media exposure produce more anxiety, and if so, what mechanisms mediate this relationship? Drawing on cognitive stress appraisal theory (Lazarus and Folkman, 1984), social comparison theory (Festinger, 1954), and the Displacement Hypothesis (Kraut et al., 1998), the paper evaluates passive and active exposure distinctions, platform-specific exposure profiles, content valence effects, and individual vulnerability moderators including attachment style, neuroticism, and prior mental health history. The paper provides a critical review of screen time as an exposure metric, arguing that duration-based measures inadequately capture the psychologically relevant dimensions of social media contact. In their place, the paper proposes a Digital Biomarker Battery (DBB) for social media anxiety assessment integrating: (1) ecological momentary assessment of state anxiety correlated with passive smartphone usage logs, (2) heart rate variability measured via consumer wearables as a physiological anxiety index, (3) linguistic markers of anxious cognition extracted from users’ own social media posts using validated NLP models, and (4) social comparison frequency as a behavioral engagement metric derived from interaction logs. Meta-analytic evidence from Yoon et al. (2019; k = 13, N = 21,006) is critically evaluated alongside the Orben and Przybylski (2019) specification curve analysis demonstrating that effect size estimates for digital technology-wellbeing associations vary dramatically depending on analytic choices, from r = -.15 to r = .10. A research agenda centered on pre-registered, multi-method, longitudinal designs is proposed.
Keywords: social media anxiety; digital biomarkers; ecological momentary assessment; dose-response; screen time; social comparison; mental health assessment; passive sensing.
