Submitted by Edward Grierson on Wed, 05/03/2025 - 09:29
Jacob Verrey, a PhD candidate at the Institute of Criminology, has designed a new methodology for collecting data on fatal shootings in the USA. He has now detailed his methodology in a paper titled ‘DM-FS: A Comprehensive Database on Death-Modulated Fatal Shootings’, published in Nature Scientific Data.
Fatal shootings in the USA have garnered more attention by the media in recent years. This has drawn attention to its prevalence in the country. From 2015 to 2019 in the USA, police officers shot 4,195 civilians, while civilians shot 256 officers. These deaths generated discussion about the relationships between US citizens and law enforcement. In particular, the death of George Floyd by a police officer in 2020 led to increased scrutiny of police-civilian relations.
Floyd’s death made Verrey interested in studying the relationship between police and civilians, which led him to investigating fatal police-civilian encounters. However, he found several underlying issues with the available data.
His research combined 1 database on officer deaths with 3 on civilian deaths, which together covered 2015 to 2020. Officer death data was assembled using FBI data, or data from the primary United States federal law enforcement agency. Meanwhile, civilian death data was obtained from 3 crowdsourced databases. However, one shared characteristic in all of these databases was that they were unidirectional. Every database focused on one aspect of the officer-civilian dynamic, which limited their reliability as sources.
“It's either ‘civilians shoot officer’, which is an officer safety perspective. Or it’s ‘police shoot civilians’, the typical fatal shooting perspective,” Verrey explains. “But this stuff is really multifaceted, and I wanted to take a balanced approach.” A unidirectional approach also created political partisanship among the groups collecting the data. While Verrey applauds them for assembling data that others can use, he also stresses that there are inevitable issues with bias. He argues that some of the datasets may be hosted by organisations with political ideologies. With this in mind, researchers should be careful about conducting research on data that might be influenced by these ideologies.
“What would be really interesting would be if we took a bi-directional approach,” he says. “To understand how the death of an officer affects the number of civilians others fatally shoot each year, and vice versa. It wouldn’t be unreasonable to think that one death, one way, will modulate deaths in the other.”
Verrey therefore began his research by designing a more impartial system of data measurement. Called ‘Death-Modulated Fatal Shootings’ (DM-FS), this bidirectional system collated deaths from all 4 databases, allowing for direct comparisons.
He then categorised the combined data into DM-FS Officers and DM-FS Civilians, which allowed the death rates to be standardised. Standardisation involves using individual deaths as baselines for broader trends. One example is a civilian shooting an officer. The number of deaths by shootings within the dead individual’s law enforcement agency are measured before and after this event. Standardisation can therefore provide more clarity on responses to fatal shootings. In particular, Verrey was interested in whether the shooting of one civilian or police officer can trigger more in response.
“No matter what you find with this database, it’ll be fascinating. Let's say that you find that a year after an officer's death, fatal shootings increase. We could create a warning system,” he suggests, “such that, if an officer dies, we could do an intervention. We would target [specific] agencies and give them training that can prevent future fatal shootings. That is a fascinating smoking gun with a potentially life-saving intervention, if it's what we find.”
“The other interesting finding would be that, after an officer dies, officers do not engage in more fatal shootings in subsequent years. That would also be fascinating, because it means that even though your colleague was killed by a civilian, other officers do not engage in more fatal shootings. That is a pretty reassuring vote of confidence for the police.”
Verrey was quick to thank Professors Lawrence Sherman and Justice Tankebe for providing the mentorship that gave rise to this line of reasoning.
Going forward, he hopes that researchers will utilise the DM-FS methodology. Their results could pose further questions about the degree to which other factors may influence fatal shootings.
Verrey gives location as a potential example. “Small towns tend to be very different from big cities, both with respect to the civilians in them and with respect to law enforcement. We know that small towns experience more fatal shootings than big cities,” he explains. “But what if small towns also respond to fatal shootings differently too? Maybe small-town police face an increased danger after the death of a civilian whereas big city police do not. DM-FS lays the foundation by which this type of modulating-factor research can be conducted.”
DM-FS can also be used to boost the validity of research findings. Another issue with existing fatal shooting research is that their findings tend to draw important measurements from one database alone. By contrast, researchers using DM-FS do not rely on measurements from only one database. They can filter DM-FS so it only displays these measurements from multiple databases at once.
“Say, for example, you want to study fatal shootings based on deaths that are recorded in the Washington Post database,” Verrey explains. “You can do that and get your first research finding. Then, conduct the same study using deaths as they are recorded in the Fatal Encounters database. You do that and get your second research finding. Compare the findings, see if they hold, and if so, your findings become a lot stronger than they otherwise would have been if you relied on one database alone.”
Verrey is optimistic that DM-FS can provide a framework for future studies that engage with multiple databases. In turn, it could assess the veracity of these sources. Doing so would ensure a more nuanced outlook on fatal shootings, both in the USA and elsewhere.
“It’s mesmerising, for lack of a better word,” he says. “Research drawing from this database directly impacts the real world, and it could be used to save a life.”