Black Swan Shootings: A Model for Predicting the Worst of the Worst Mass Shootings

Matthew D'Anna

Advisor: Christopher Koper, PhD, Department of Criminology, Law and Society

Committee Members: Cynthia Lum, Catherine Gallagher, Guoqing Diao

Online Location, Online
May 08, 2020, 10:00 AM to 12:00 PM


Since the late 1990s, mass shootings, by arguably any definition, have increased in frequency and lethality in the United States at an alarming and largely unprecedented rate (Lankford and Silver, 2019; Krouse and Richardson, 2015). Most academic studies on mass shootings are descriptive and offender-centric, focusing on biographic traits, individual causes, and poor methods for identification of early warning signs. Thus, there are often little-to-no opportunities for predicting future mass violence. This study takes a different approach in multiple ways. First, rather than analyzing all mass shootings, this study identifies ‘Black Swan Shootings’, defined here as the rarest and most violent mass shootings in recent history. Second, this study focuses on location-based patterns across community-level measures for contagion, safety, demographics, stability, and weapons. Third, rather than predicting offenders, this study develops a spatial threat assessment for identifying high-risk attack locations and offender residences. Thus, there are significant relationships between U.S. counties experiencing Black Swan attacks and violent crime, population density, race, sex, poverty, drug and alcohol overdoses, suicide, and firearm availability. Further, there are significant relationships for Black Swan offender residence counties and violent crime, population density, race, sex, drug and alcohol overdoses, and firearm availability. These variables are used in a threat assessment model to attempt to identify the annual count and location of Black Swan Shootings. The model’s performance is evaluated, with implications for identification of counties at the highest risk for a future attack.