Posted on

Validity of Gun Violence Archive as Data Source for Community Firearm Violence Assessed in US Cities

Study Evaluates Gun Violence Archive’s Reliability in Tracking Firearm Violence Events in Philadelphia, New York, Chicago, and Cincinnati

A new study has assessed the reliability of the Gun Violence Archive (GVA) as a data source for community firearm violence, particularly firearm homicide and nonfatal shootings resulting from interpersonal violence. This evaluation, conducted in four major U.S. cities over a six-year period, sheds light on the strengths and limitations of GVA data.

Key Findings: The study, which examined data from Philadelphia, New York, Chicago, and Cincinnati from 2015 to 2020, found that the GVA displayed an overall sensitivity of 81.1%. In other words, it correctly identified 81.1% of community firearm violence shooting events, but with some noteworthy variations.

GVA Strengths and Weaknesses: The findings suggest that the Gun Violence Archive has certain unique advantages, such as spatial resolution, timeliness, and geographic coverage. It is particularly useful for researchers in large cities. However, caution is advised when conducting a granular examination of epidemiology, as the data source exhibits systematic biases.

Study’s Approach: To validate the GVA’s data, the study compared GVA records with publicly available police department data, which served as the reference standard. This cross-sectional analysis focused on the event-level and person-level validity of the GVA. It aimed to assess how accurately the GVA identifies community firearm violence events and the characteristics of individuals injured in these events.

Changes Over Time: The study revealed variations in sensitivity over the years, which is essential information for studies examining trends in shootings. Sensitivity improved over time, but this improvement differed among the cities evaluated. The sensitivity was lower in 2016 and increased to excellent levels by 2019 in the major cities. Cincinnati, a midsize city, had lower sensitivity throughout the study period.

Systematic Missingness: The GVA demonstrated a systematic bias in terms of the type of shootings it accurately documented. Fatal shootings and incidents involving women and children were less likely to be missing from the GVA, suggesting that the GVA’s data collection method may reflect media biases in covering community firearm violence.

Recommendations for Research: The study recommends that future research consider the limitations and potential biases of the GVA when using its data. Caution is advised when analyzing GVA data from earlier years and conducting time trend analyses. Additionally, researchers are encouraged to assess the GVA’s validity in other cities to ascertain its reliability as a national source of community firearm violence data.

Conclusion: The study highlights the GVA’s role as a valuable resource for researchers studying community firearm violence in major U.S. cities. However, an awareness of its limitations and biases is essential for accurate interpretation and analysis, ensuring that findings are not inadvertently skewed due to data biases and missingness.

Source: JAMA Network Open Journal