Smartphones and teen mental health: what Twenge's case actually rests on

Jean Twenge's iGen (2017) and Jonathan Haidt's The Anxious Generation (2024) advance a specific causal claim: smartphones and social media caused the post-2010 collapse in adolescent mental health. The argument has been politically consequential. It has also been one of the most contested empirical claims in current social science.

The honest summary is more nuanced than either the strong popular version or the dismissive academic counter.

1. What's not in doubt

Several things are well-established and uncontroversial:

• Self-reported depression, anxiety, and suicidality among American adolescents rose substantially beginning around 2011-2012, particularly among girls
• The same period coincides with the saturation of smartphones in adolescent life (US smartphone adoption among teens crossed 50% in 2012, 90% by 2017)
• Adolescents who use social media heavily report worse mental health, on average, than light users
• Heavy nighttime smartphone use disrupts sleep, which independently affects mental health

These four findings are robust. The question is how to causally relate them.

2. The correlational evidence

Cross-sectional studies consistently find a small-to-moderate correlation between social media use and adolescent mental health problems, around r = 0.10 to 0.20. The correlation is larger for girls than boys. It is larger for emotional outcomes (depression, anxiety) than for behavioral outcomes (substance use, aggression).

Orben and Przybylski's 2019 analysis emphasized that this correlation is roughly the same size as the correlation between eating potatoes and adolescent mental health, or between wearing glasses and mental health. The finding is real but small (Orben & Przybylski, 2019).

Twenge's counter: at the population level, even small individual effects scale to substantial public-health impact when distributed across hundreds of millions of users.

3. The longitudinal and experimental evidence

This is where the argument gets less certain. Several longitudinal studies have tracked individual changes:

• Heavy social media users show worse subsequent mental health than light users, after controlling for baseline (Twenge et al., 2018)
• But also: adolescents in worse mental health subsequently increase social media use, suggesting reverse causation
• Bidirectional effects appear in the cleanest longitudinal designs

A small number of experimental studies — randomly reducing teen social media use for weeks — have found short-term mental health benefits. Effect sizes are modest (Allcott et al., 2020).

4. The cross-national pattern

If smartphones caused the decline, the trajectory should be similar across smartphone-saturated countries. The picture is mixed: rises in adolescent depression are visible in the US, UK, Canada, Australia, and parts of Northern Europe — but not consistently in Asian countries with similar smartphone penetration. The cross-national heterogeneity is hard to reconcile with a simple smartphone-causes-depression story.

This suggests the smartphone effect, if real, operates through specific cultural conditions — social comparison, school structure, parenting norms — that vary across societies.

5. The honest summary

Twenge's strong claim — that smartphones caused the adolescent mental health crisis — overshoots the evidence in confidence and specificity. The weaker claim — that heavy social media use is one of several factors contributing to worsening adolescent mental health in specific cultural contexts — is reasonably well-supported.

The policy implications are similar either way. Reducing heavy adolescent smartphone use, particularly at night and during school, is supported by what we know. Eliminating smartphones from adolescent life is supported by less than what proponents claim.

The empirical answer about whether smartphones are the cause may not arrive in time for the policy decisions that need to be made. This is uncomfortable but realistic.