Firearm Purchase Behavior and Subsequent Adverse Events, United States, 1985-2018

The primary purpose of this project was to develop and test a new large-scale approach to threat assessment that relies on objective data regarding firearm purchases. Specifically, to analyze firearm transaction records in California to better understand the firearm purchasing patterns of mass shooters and perpetrators of firearm-related crimes and to build risk prediction models to help identify individuals who might be at extreme risk for committing such crimes in the future.

The objectives of the work included:

• Objective 1: To determine whether there were unusual pre-event firearm purchasing patterns among known mass shooters with a record of purchase in California (1985-2018), as compared to the general population of registered firearm purchasers in the state of California.
• Objective 2. To determine whether perpetrators of homicide, robbery, or aggravated assault in California with a record of purchase (1985-2018) have distinct pre-event firearm purchasing patterns as compared to the general population of registered firearm purchasers in the state.
• Objective 3: To use machine learning methods to forecast who is at an elevated risk of involvement in a mass shooting based on features generated from firearm transaction records, along with criminal history records, and other key individual and community characteristics available in the administrative data.
• Objective 4: To use the same machine learning methods and predictor variables to forecast who is at elevated risk of involvement in other serious firearm violence (homicide, robbery, or aggravated assault with a firearm).