Social Media and Adolescent Mental Health: The Causation Debate Remains Unresolved

Few topics in contemporary psychology generate as much heat — and as little clarity — as the relationship between social media use and adolescent mental health. Adolescent depression and anxiety have increased substantially over the past decade, and social media use has increased over the same period. But correlation is not causation, and the research community remains divided on whether social media is a primary driver of the adolescent mental health crisis, a contributing factor among many, or largely irrelevant. Recent reviews and empirical studies offer important new evidence, but the fundamental causal question remains open.

The Research Landscape

The Case for Harm

Burgess (2025), in a narrative review published in Creative Nursing, synthesizes research from 2016 to 2024 linking social media use to increased depression, anxiety, and suicidal ideation among adolescents aged 13-19. The review identifies several mechanisms through which social media may harm mental health:

• Social comparison and self-esteem erosion: Constant exposure to curated, idealized portrayals of others' lives leads to negative social comparisons.
• Cybervictimization: Cyberbullying operates continuously, without the physical-world limits of school-based bullying.
• Fear of missing out (FOMO): Awareness of social events one is not participating in generates anxiety and exclusion feelings.
• Contagion phenomenon: Exposure to self-harm and suicidal content on social media may normalize these behaviors.
• Addiction mechanisms: Variable-ratio reinforcement schedules (notifications, likes, comments) exploit the same reward pathways as behavioral addictions.

The Nuanced View: Not All Use Is Equal

Agyapong-Opoku, Agyapong-Opoku, and Greenshaw (2025), in a scoping review of reviews analyzed 43 systematic reviews, meta-analyses, and narrative reviews published between 2020 and 2024. Their findings complicate the simple "social media harms teens" narrative:

• Problematic use and passive consumption are most strongly associated with harm. Active social media use — posting, commenting, messaging friends — shows weaker associations with negative outcomes than passive scrolling.
• Some studies report positive effects, including enhanced social support, reduced isolation, and access to mental health resources — particularly for LGBTQ+ youth and other marginalized populations.
• COVID-19 effects were mixed: During the pandemic, social media served as a critical social lifeline for isolated adolescents, but also intensified exposure to distressing content.
• The relationship is "highly individualistic": Moderating factors — personality, existing mental health conditions, family support, digital literacy — shape whether social media use is harmful or beneficial.

Causal Evidence Remains Thin

Fruehwirth, Weng, and Perreira (2024) provide one of the stronger quasi-experimental studies on causal effects. Using longitudinal data from college students collected before and during the COVID-19 pandemic, they exploit temporal variation to estimate causal effects:

• Small, insignificant effects at 4 months into the pandemic during social distancing.
• Large, significant negative effects at 18 months when colleges resumed normal operations.
• The negative effects were concentrated among socially isolated students — those lacking in-person social support.
• Social support and resilience were protective factors: Students with strong social networks were largely unaffected by increased social media use.

This study is notable because it goes beyond cross-sectional associations to provide evidence of causal pathways — but the sample is college students, not younger adolescents, and the pandemic context limits generalizability.

Peer Relationships as a Confound

Stuke, Schlack, and Erhart (2025) complicate the narrative further with their machine learning analysis of two large German cohort studies. Using interpretable causal inference methods, they find that peer relationship quality is a more direct predictor of adolescent mental health than screen time. Their provocative claim: the adolescent mental health crisis is driven primarily by deteriorating peer relationships, and social media is, at most, a mediating factor — the tool through which peer dynamics play out, not the root cause.

Critical Analysis: Claims and Evidence

<

Claim	Evidence	Verdict
Social media use causes adolescent depression/anxiety	Multiple reviews report association, few establish causation	⚠️ Association strong, causation unproven
Passive consumption is more harmful than active use	Agyapong-Opoku et al. scoping review, 43 reviews analyzed	✅ Consistent finding across multiple reviews
Social media has protective effects for some groups	Agyapong-Opoku et al. — LGBTQ+ youth, isolated populations	✅ Supported — but effect size unclear
Social isolation moderates harm from social media	Fruehwirth et al. longitudinal quasi-experiment	✅ Supported — strong design
Peer relationships matter more than screen time	Stuke et al. causal ML analysis, two cohorts	⚠️ Plausible — but single study, needs replication

Open Questions and Future Directions

Experimental evidence is still scarce. The gold standard — randomized assignment to reduce or eliminate social media use — has been attempted in only a handful of small studies. Larger, longer experiments are needed to establish causal direction.

Platform matters. Most research treats "social media" as a monolith. TikTok, Instagram, Snapchat, YouTube, and Discord operate through different mechanisms and likely have different effects. Platform-specific research is essential.

Age-dependent effects. The Fruehwirth et al. study used college students. Effects on 10-14 year-olds — who are undergoing different neurodevelopmental processes — may be qualitatively different.

Algorithmic amplification. The role of recommendation algorithms in shaping adolescent exposure to harmful content has received little empirical attention, despite being central to the public policy debate.

Policy without certainty. Legislators in multiple countries are implementing age restrictions, notification bans, and other regulations — often ahead of the evidence. The research community bears responsibility for providing clearer, more actionable findings.

What This Means for the Field

The honest assessment of the current evidence is this: social media use is associated with worse mental health outcomes for some adolescents, under some conditions, through some mechanisms — but the evidence for a simple causal relationship is weaker than public discourse suggests. The most robust finding is that how adolescents use social media matters more than whether they use it, and that individual vulnerability factors (social isolation, existing mental health conditions, low digital literacy) moderate the relationship substantially.

For clinicians, the implication is that "quit social media" is too blunt a recommendation. For policymakers, the implication is that evidence-based regulation requires more nuance than blanket bans. For researchers, the priority is rigorous experimental and longitudinal work that can disentangle correlation from causation.

Explore related work through ORAA ResearchBrain.

Few topics in contemporary psychology generate as much heat — and as little clarity — as the relationship between social media use and adolescent mental health. Adolescent depression and anxiety have increased substantially over the past decade, and social media use has increased over the same period. But correlation is not causation, and the research community remains divided on whether social media is a primary driver of the adolescent mental health crisis, a contributing factor among many, or largely irrelevant. Recent reviews and empirical studies offer important new evidence, but the fundamental causal question remains open.

The Research Landscape

The Case for Harm

Burgess (2025), in a narrative review published in Creative Nursing, synthesizes research from 2016 to 2024 linking social media use to increased depression, anxiety, and suicidal ideation among adolescents aged 13-19. The review identifies several mechanisms through which social media may harm mental health:

• Social comparison and self-esteem erosion: Constant exposure to curated, idealized portrayals of others' lives leads to negative social comparisons.
• Cybervictimization: Cyberbullying operates continuously, without the physical-world limits of school-based bullying.
• Fear of missing out (FOMO): Awareness of social events one is not participating in generates anxiety and exclusion feelings.
• Contagion phenomenon: Exposure to self-harm and suicidal content on social media may normalize these behaviors.
• Addiction mechanisms: Variable-ratio reinforcement schedules (notifications, likes, comments) exploit the same reward pathways as behavioral addictions.

The Nuanced View: Not All Use Is Equal

Agyapong-Opoku, Agyapong-Opoku, and Greenshaw (2025), in a scoping review of reviews analyzed 43 systematic reviews, meta-analyses, and narrative reviews published between 2020 and 2024. Their findings complicate the simple "social media harms teens" narrative:

• Problematic use and passive consumption are most strongly associated with harm. Active social media use — posting, commenting, messaging friends — shows weaker associations with negative outcomes than passive scrolling.
• Some studies report positive effects, including enhanced social support, reduced isolation, and access to mental health resources — particularly for LGBTQ+ youth and other marginalized populations.
• COVID-19 effects were mixed: During the pandemic, social media served as a critical social lifeline for isolated adolescents, but also intensified exposure to distressing content.
• The relationship is "highly individualistic": Moderating factors — personality, existing mental health conditions, family support, digital literacy — shape whether social media use is harmful or beneficial.

Causal Evidence Remains Thin

Fruehwirth, Weng, and Perreira (2024) provide one of the stronger quasi-experimental studies on causal effects. Using longitudinal data from college students collected before and during the COVID-19 pandemic, they exploit temporal variation to estimate causal effects:

• Small, insignificant effects at 4 months into the pandemic during social distancing.
• Large, significant negative effects at 18 months when colleges resumed normal operations.
• The negative effects were concentrated among socially isolated students — those lacking in-person social support.
• Social support and resilience were protective factors: Students with strong social networks were largely unaffected by increased social media use.

This study is notable because it goes beyond cross-sectional associations to provide evidence of causal pathways — but the sample is college students, not younger adolescents, and the pandemic context limits generalizability.

Peer Relationships as a Confound

Stuke, Schlack, and Erhart (2025) complicate the narrative further with their machine learning analysis of two large German cohort studies. Using interpretable causal inference methods, they find that peer relationship quality is a more direct predictor of adolescent mental health than screen time. Their provocative claim: the adolescent mental health crisis is driven primarily by deteriorating peer relationships, and social media is, at most, a mediating factor — the tool through which peer dynamics play out, not the root cause.

Critical Analysis: Claims and Evidence

<

Claim	Evidence	Verdict
Social media use causes adolescent depression/anxiety	Multiple reviews report association, few establish causation	⚠️ Association strong, causation unproven
Passive consumption is more harmful than active use	Agyapong-Opoku et al. scoping review, 43 reviews analyzed	✅ Consistent finding across multiple reviews
Social media has protective effects for some groups	Agyapong-Opoku et al. — LGBTQ+ youth, isolated populations	✅ Supported — but effect size unclear
Social isolation moderates harm from social media	Fruehwirth et al. longitudinal quasi-experiment	✅ Supported — strong design
Peer relationships matter more than screen time	Stuke et al. causal ML analysis, two cohorts	⚠️ Plausible — but single study, needs replication

Open Questions and Future Directions

Experimental evidence is still scarce. The gold standard — randomized assignment to reduce or eliminate social media use — has been attempted in only a handful of small studies. Larger, longer experiments are needed to establish causal direction.

Platform matters. Most research treats "social media" as a monolith. TikTok, Instagram, Snapchat, YouTube, and Discord operate through different mechanisms and likely have different effects. Platform-specific research is essential.

Age-dependent effects. The Fruehwirth et al. study used college students. Effects on 10-14 year-olds — who are undergoing different neurodevelopmental processes — may be qualitatively different.

Algorithmic amplification. The role of recommendation algorithms in shaping adolescent exposure to harmful content has received little empirical attention, despite being central to the public policy debate.

Policy without certainty. Legislators in multiple countries are implementing age restrictions, notification bans, and other regulations — often ahead of the evidence. The research community bears responsibility for providing clearer, more actionable findings.

What This Means for the Field

The honest assessment of the current evidence is this: social media use is associated with worse mental health outcomes for some adolescents, under some conditions, through some mechanisms — but the evidence for a simple causal relationship is weaker than public discourse suggests. The most robust finding is that how adolescents use social media matters more than whether they use it, and that individual vulnerability factors (social isolation, existing mental health conditions, low digital literacy) moderate the relationship substantially.

For clinicians, the implication is that "quit social media" is too blunt a recommendation. For policymakers, the implication is that evidence-based regulation requires more nuance than blanket bans. For researchers, the priority is rigorous experimental and longitudinal work that can disentangle correlation from causation.

Explore related work through ORAA ResearchBrain.