Sarah Oliver

As a research assistant at Boston College, I co-authored a chapter in Gun Violence: The Statistical Issues investigating how the COVID-19 pandemic impacted the contagiousness of mass shootings in the United States. Using Hawkes process models, we found that the reproduction number of mass shootings nearly doubled after March 2020, a result consistent across three national databases. The book will soon be published to the Gun Violence Archive.

• Used Python and STAN (a probabilistic programming language) for Bayesian estimation of model parameters
• Simulated a Hawkes self-excitatory point process to model the contagion effect of mass shootings over time
• Estimated a discontinuity in the reproduction number before and after March 2020, finding it increased from 0.225 to 0.417 in the GVA dataset
• Analyzed three national datasets (Gun Violence Archive, Mother Jones, USA Today/AP/Northeastern) to test robustness across different mass shooting definitions

• The reproduction number of mass shootings nearly doubled after the start of the COVID-19 pandemic across all three datasets
• Both the baseline rate and contagiousness of mass shootings increased after March 2020
• The increase was robust to sensitivity analysis using alternative cutoff dates
• Results are consistent with the hypothesis that pandemic-era changes in online activity and routine activities amplified contagion dynamics