Social Media & Youth Mental Health

The underlying trend is real. The CDC’s 2023 Youth Risk Behavior Survey found 39.7% of U.S. high school students reported persistent sadness or hopelessness, 20.4% had seriously considered suicide, and 9.5% had attempted suicide. The sex gap is large: among girls, 52.6% reported persistent sadness versus 27.7% of boys. These problems were worsening before COVID, not just during it.

Source: CDC YRBS 2023 — nationally representative school-based survey of ~20,000 students.

UK NHS Digital reports probable mental disorder in ages 8–16 rose from 12.1% (2017) to 20.3% (2023); for ages 17–19 it reached 23.3%. OECD and UNICEF data show worsening adolescent well-being in many high-income countries between the mid-2010s and early 2020s. WHO Europe found problematic social media use rising from 7% (2018) to 11% (2022). But the pattern is not universal: some countries did not show the same deterioration, weakening any claim that one platform alone explains everything everywhere.

Source: CDC YRBS 2013–2023. Slight improvement 2021–2023 (girls: 57%→53%) may reflect post-peak stabilization or survey-mode effects.

Most studies are cross-sectional or longitudinal with self-reported screen time. These show consistent but small associations. Gabrielle et al. (2024 meta-analysis, 45 studies): r=0.12 for depression, r=0.10 for anxiety, r=0.15 for loneliness. Orben & Przybylski (2019) found digital technology explained ≤0.4% of variance in well-being across three large datasets — comparable in magnitude to wearing glasses. Self-reported screen time is imprecise: passive device measures show only moderate correlation with self-report (r=0.18–0.64).

ABCD within-person panel (Nagata et al. 2025): Following 11,876 U.S. children, increases in a teen’s own social-media use above their usual level predicted later increases in depressive symptoms (β≈0.07–0.09). The reverse path (depression → later use) was not significant. Small effect, but prospective and within-person.

Braghieri, Levy & Makarin (AER 2022): Facebook’s staggered rollout across U.S. colleges increased depression by 0.085 SD (~2 percentage points) and anxiety by ~12%. Strongest quasi-experimental evidence, but limited to college students on early Facebook — not today’s teens on TikTok/Instagram.

Reduction RCTs (meta-analysis 2025): Small but consistent benefits from curtailing use (Hedges’ g≈0.19–0.28 for depression/anxiety/well-being). Most studies are on adults, not adolescents. SMART Schools UK study (2025) found school phone bans reduced in-school use but did not improve overall mental health.

Jonathan Haidt (NYU Stern, social psychologist): Core claim — smartphone-based social media is a major causal driver of the post-2012 crisis, especially for girls. Relies on trend timing, sex differences, mechanisms, quasi-experiments, and precautionary logic. Strongest critique: Often infers more causality than underlying designs support; extrapolates from college/adult studies.

Candice Odgers (UC Irvine, developmental psychologist): Core claim — evidence for a broad, strong causal story is weak; most associations are small or mixed. Stresses reverse causality, confounding, and measurement error. Strongest critique: May underweight subgroup harms and dismiss population-level consequences of small effects.

Amy Orben & Andrew Przybylski (Cambridge/Oxford): Average associations tiny (≤0.4% variance); analytic flexibility inflates apparent effects. Jean Twenge (SDSU): Stronger negative effects for girls; heavy use doubles risk. Mitch Prinstein (APA): Context-dependent — effects vary by individual and use patterns.

Sex differences are robust and large. YRBS: girls’ persistent sadness rose from 39% to 53% (2013–2023) vs. boys’ 21% to 28%. Meta-analyses find effect sizes for social media and depression are 2–3× larger for girls (Twenge et al. 2022: β=−0.11 to −0.24 for girls vs. −0.03 to −0.06 for boys). WHO HBSC (2022) found problematic social media use at 13% of girls vs. 9% of boys.

Why girls may be more affected: Girls use more image-based, appearance-focused platforms (Instagram, TikTok). Greater exposure to social comparison, body-image pressures, cyberbullying, and relational aggression. Girls are more prone to internalizing disorders. Orben et al. (2022) found developmental “windows of sensitivity” with small negative effects earlier for girls (ages 11–13).

Source: Twenge et al. (2022) specification curve analysis. β range for girls −0.11 to −0.24; boys −0.03 to −0.06. Midpoints shown.

Sleep displacement has the strongest evidence: late-night use and notifications disrupt sleep, which mediates negative mental health effects. Social comparison on appearance-focused, passive-scrolling platforms increases body dissatisfaction and anxiety, especially among girls. Cyberbullying is linked to increased depression and suicidality. Compulsive/problematic use (addiction-like patterns in ~10–20%) shows the strongest associations — Shannon et al. (2022) found r=0.27 for depression and r=0.35 for anxiety, much larger than ordinary time-spent correlations.

Social support and belonging: Online communities can reduce isolation, especially for marginalized or geographically isolated youth. Identity development: LGBTQ+ youth and teens in restrictive environments often find affirmation and community online that is unavailable offline. Access to information and help: Some teens use social media to find mental health resources and peer support. APA and the National Academies both note that digital spaces can provide real opportunities alongside risks.

The 2023 advisory is explicitly precautionary, not causal. It states: “We do not yet have enough evidence to determine if social media is sufficiently safe for children and adolescents.” It cites concerning correlational findings but explicitly notes gaps in pathways, content effects, and subgroup data. It recommends safety standards, design changes, and research transparency — not bans. Media headlines often compressed this into “social media harms youth mental health,” overstating the advisory’s actual conclusions.