The Dog That Didn't Bark: School Suicide Prevention Training Mandates and Population Mortality

Abstract

Over 30 US states have mandated suicide prevention gatekeeper training for school personnel, yet no causal evidence exists on whether these laws reduce suicide mortality. I exploit staggered adoption of mandatory training laws across 25 states between 2007 and 2017 using the heterogeneity-robust difference-in-differences estimator. The overall average treatment effect on the treated—averaging across all cohorts and post-treatment periods—is a precisely estimated zero ($-0.014$ per 100,000, $p = 0.96$). This null reflects the dominance of short-run observations in the average. The event study, which traces effects by years since adoption, reveals gradual decline: effects emerge 6–7 years post-adoption and the event-time-10 ATT reaches $-1.78$ per 100,000 (95% CI: $[-2.49, -1.06]$, $p < 0.001$). Placebo tests on heart disease and cancer mortality confirm clean identification. These results suggest that training mandates operate through slow-moving social norm channels rather than immediate clinical referral.