This study assessed the sensitivity of Gunshot Detection Technology (GDT) relative to Calls for Service.
Existing crime data sources have biases that do not present a complete picture of gun-related crime. GDT may provide a new metric for firearm crimes; however, few studies have assessed the accuracy of GDT and its relationship to other measures of violence. GDT and gun crime-related Calls for Service in Washington, DC during 2010 were studied. Using temporal comparisons for month, day of year, weekday, and hour of the day, spatial comparisons on a quarter-mile square grid, and a Poisson-Lognormal-CAR spatial regression model on a combined grid by time period dataset, this study examined the sensitivity of GDT activations relative to gunshot-related calls for service. The results showed that relative GDT sensitivity changed by time and by space. In particular, the relative sensitivity of GDT was much stronger in the evening and at nighttime than in the daytime. In terms of spatial variation, the study found that GDT sensitivity decreased with distance from the nearest zone centroid. In addition, there were two small geographic areas in the study area in which the relative GDT sensitivity was lower than expected. GDT systems identify the frequency and locational accuracy of gunshot incidents, particularly at nighttime. This technology has the potential to improve data collection on gun use and violence and produce more accurate representations if the temporal and distance limitations of the technology are understood. GDT may improve gun detection and, thereby, improve police operations and public support for police. (publisher abstract modified)