Social Media Use and Mental Health in Adolescents and Young Adults: A Critical Review of Recent Evidence

• General social media use (SMU): Ordinary engagement with social networking platforms, with no assumption of harm.

• Problematic social media use (PSMU): A pattern of use marked by impaired control or interference with daily functioning. It is not the same as “addiction” unless an addiction measure was validated.

• Problematic internet use (PIU): A broader pattern covering all online activities, not only social networking.

• Narrative review: A selective, descriptive literature review. Unlike a systematic review, it does not use a documented, reproducible search.

• Systematic review / meta-analysis / umbrella review: A systematic review collects studies by a structured method. A meta-analysis statistically pools their results. An umbrella review summarizes several existing reviews.

• Correlation (r): A number showing how strongly two things vary together. It shows association, not cause.

• Observational / cross-sectional / longitudinal / experimental designs: Observational studies watch people without intervening. Cross-sectional studies measure everyone once. Longitudinal studies follow people over time. Experimental studies assign people to conditions, and only these can strongly support cause and effect.

• Fear of missing out (FoMO): Anxiety about being left out of others’ experiences.

• Cyberbullying (victimization/perpetration): Harassment online. Victimization means being targeted; perpetration means doing the targeting.

• Social comparison: Judging yourself against others, here often against idealized online images.

• Heterogeneity: Differences between studies in methods, samples, and results.

1. How is social media use associated with depressive and anxiety symptoms in adolescents and young adults?

2. How do associations differ between general use and problematic use?

3. What role do loneliness, cyberbullying, and social comparison play?

4. Can online communities and social support offer benefits?

5. Who may be more vulnerable?

The central argument is that the quality, context, and consequences of use are more informative than time spent alone.

• Review approach: This is a critical narrative review of literature cited in the source manuscript. It focuses on systematic reviews, meta-analyses, and umbrella reviews. No reproducible database search, screening criteria, or risk-of-bias assessment was documented, so selection bias is possible. The author calls it a structured overview of selected evidence, not a systematic review.

• General vs. problematic use: General social media use showed small, variable associations with mental health outcomes. Problematic use showed more consistent associations with depression, anxiety, sleep difficulties, and well-being (Ahmed et al., 2024).

• Size of the associations: A meta-analysis of 18 studies with 9,269 young participants found moderate positive correlations between PSMU and depressive symptoms (r = .273), anxiety (r = .348), and stress (r = .313) (Shannon et al., 2022). The review stresses these show association, not diagnosis or cause.

• Anxiety: Possible contributors include constant notification monitoring, concern about social feedback, and FoMO, but no single mechanism is established. A meta-analysis of 37 studies with over 36,000 young people linked problematic internet use to social anxiety (Ding et al., 2023). Because that evidence concerns the internet broadly, it is adjacent context, not a direct estimate for social media.

• Depression: Associations are more consistent when problematic use is measured than when frequency or duration alone is. Cross-sectional data cannot show whether problematic use comes before symptoms, follows them, or both stem from other factors. Phrases like “causes depression” should be avoided without longitudinal or experimental support.

• Loneliness: In a 2025 UK cohort, total time online was associated with loneliness but time on social media specifically was not. Problematic use and online victimization were more informative correlates (Matthews et al., 2025). Loneliness also reflects wider social and economic conditions. The Surgeon General’s smoking comparison concerns social disconnection generally, not social media harm.

• Cyberbullying and comparison: Cyberbullying is a plausible pathway to distress, though the cited review is older (2015) and its findings are treated as foundational. Social comparison is plausible, but popular commentary is not treated as empirical evidence.

• Benefits: Social media can offer peer support, identity-affirming communities, health information, and self-expression. These may especially help people who face stigma or limited access to in-person support. Among university students, perceived social support through social media was associated with life satisfaction, whereas problematic patterns were linked to less favorable outcomes.

• Vulnerability: An umbrella review found mixed evidence that varies by individual and context (Valkenburg et al., 2022). Age alone does not determine vulnerability. Gender, prior mental health difficulties, family and peer relationships, socioeconomic conditions, and platform features may all matter.

Overall interpretation: Social media and mental health are related in complex, context-dependent ways. The evidence supports neither “uniformly harmful” nor “uniformly beneficial.” Because most findings are observational, they should not be described as established causal effects.

Limitations acknowledged by the author:

• The underlying literature is mostly observational and relies on self-reported use and symptoms.

• Definitions of general use, problematic use, and addiction-like behavior are inconsistent.

• Platforms change quickly, and studies differ in ages, cultures, measures, and outcomes.

• The review is selective, with no documented search or bias assessment, and publication bias cannot be excluded.

Recommendations for practice:

• Schools and health services should strengthen digital-literacy education. It should cover recognizing problematic patterns, online harassment, sleep disruption, and harmful comparison.

• Clinicians may ask about digital experiences when distress or functional impairment makes them relevant. This should be individualized, and ordinary social media use should not be equated with disorder.

• Online communities and digital mental-health resources may complement existing services, but their effectiveness and safety should be evaluated rather than assumed.

Recommendations for research:

• Use longitudinal and, where appropriate, experimental designs.

• Distinguish duration of use from purpose and impairment.

• Use validated measures and report uncertainty.

• Use transparent, reproducible review methods.