Social Media's Like Button and the Teen Mental Health Collapse After 2012

A persistent critique of the social-media-harms thesis points to studies like Orben and Przybylski's specification curve analysis, which ran tens of thousands of analytical combinations and found effects of digital technology on wellbeing comparable to eating potatoes. The methodological rebuttal is precise: these analyses committed a double misspecification. On the exposure side, they aggregated all screen time — television, gaming, web browsing, social media — into a single variable, drowning the social media signal in irrelevant noise. On the outcome side, they used broad life-satisfaction scales where only a fraction of items tapped depression or anxiety. The analogy is apt: sending forensic evidence to a lab that runs the wrong tests, then declaring the suspect cleared.

When the analysis is properly specified — isolating heavy social media use as the exposure and depression, anxiety, self-harm, and suicide as outcomes — the effect sizes become substantial and consistent. Crucially, several of these outcomes are not vulnerable to self-report bias. Emergency department admissions for deliberate self-harm and completed suicide are objective, administratively recorded events. They show the same inflection point around 2012, the same disproportionate impact on adolescent girls, and the same pattern across Anglophone nations. The dose-response relationship further strengthens the case: light use carries no measurable risk elevation, while four-to-five-hour daily use correlates with marked increases.

The temporal specificity matters. The transformative variable is not the smartphone per se but the social feedback architecture deployed between 2009 and 2012 — like buttons, retweets, Algorithmic curation. This is when adolescent social life migrated into a system of continuous public performance and real-time quantified evaluation, and it maps precisely onto the onset of the mental health deterioration.