Comparison of rampage and non-rampage mass shootings in the U.S.: A 5-year demographic analysis

Abstract

Background: Rampage mass shootings (RMS) are a subset of mass shootings occurring in public involving random victims. Due to rarity, RMS are not well-characterized. We aimed to compare RMS and NRMS. We hypothesized that RMS and NRMS would be significantly different with respect to time/season, location, demographics, victim number/fatality rate, victims being law enforcement, and firearm characteristics.

Study design: Mass shootings (4 or more victims shot at a single event) from 2014-2018 were identified in the Gun Violence Archive (GVA). Data were collected from the public domain (e.g. news). Crude comparisons between NRMS and RMS were performed using Chi-squared or Fisher's exact tests. Parametric models of victim and perpetrator characteristics were conducted at the event level using negative binomial regression and logistic regression.

Results: There were 46 RMS and 1626 NRMS. RMS occurred most in businesses (43.5%), whereas NRMS occurred most in streets (41.1%), homes (28.6%), and bars (17.9%). RMS were more likely to occur between 6AM-6PM (OR=9.0 (4.8-16.8)). RMS had more victims per incident (23.6 vs. 4.9, RR: 4.8 (4.3,5.4)). Casualties of RMS were more likely to die (29.7% vs. 19.9%, OR: 1.7 (1.5,2.0)). RMS were more likely to have at least one police casualty (30.4% versus 1.8%, OR: 24.1 (11.6,49.9)) or police death (10.9% versus 0.6%, OR: 19.7 (6.4,60.3)). RMS had significantly greater odds that casualties were adult (OR: 1.3 (1.0,1.6)) and female (OR: 1.7 (1.4,2.1)). Deaths in RMS were more likely to be female (OR: 2.0 (1.5,2.5)) and White (OR: 8.6 (6.2,12.0) and less likely to be children (OR: 0.4 (0.2,0.8)). Perpetrators of RMS were more likely to die by suicide (34.8%), be killed by police (28.3%), or be arrested at the scene (26.1%), while more than half of perpetrators from NRMS escaped without death or apprehension (55.8%). Parametric models of perpetrator demographics indicated significant increases in the odds that a RMS shooter was White (OR: 13.9 (7.3,26.6)) or Asian (OR: 16.9 (3.7,78.4)). There was no significant difference in weapon type used (p=0.35).

Conclusion: The demographics, temporality, and location differ between RMS and NRMS, suggesting that they are dissimilar and require different preventive approaches.

Keywords: Gun violence; Mass casualty; Mass murder; Mass shooting; Rampage.