Abstract
INTRODUCTION: Evidence suggests that a non-negligible number of mass violence perpetrators experience psychological distress or crisis before the attack. Prior research comparing histories of self- and other-harm suggested that those who engaged in both face elevated risks of adverse outcomes. However, despite its potential, this framework has not been used to examine pre-attack crisis signs among mass shooters. METHODS: Using The Violence Project Database, we analyzed data from 197 U.S. mass shooters in public incidents (1966-2023). Participants were classified into harm-level groups: no-harm, single-harm (either self- or other-harm), and dual-harm (both). We conducted a latent class analysis (LCA) using eight crisis indicators from the months preceding the shootings to identify unobserved profiles. RESULTS: One-third had a history of suicidality, over half had harmed others, and 20.3 % met the criteria for dual-harm. The LCA identified four classes: (1) Depressed and agitated, (2) No-crisis, (3) Psychotic symptoms, and (4) Emotionally labile and aggressive. The Emotionally labile and aggressive class showed a significant overrepresentation of dual-harm individuals and was marked by abusive behavior, mood instability, and agitation. The No-crisis group, the largest class, showed low levels of crisis indicators, even though many members had histories of harm. CONCLUSIONS: Findings suggest that mass shooters vary in the pre-attack crisis indicators and can be classified into distinct groups. Although three-fifths of mass shooters had no pre-attack crisis indicators, a non-trivial number did exhibit identifiable signs of crisis, highlighting the heterogeneity in pre-attack crisis indicators and suggesting the need for prevention strategies that account for different profiles and needs. A harm-based trichotomy framework (no-harm, single-harm, and dual-harm) may help identify those with more acute emotional and behavioral risk.