Database provides details about officer-involved shootings and deadly incidents

When examining data regarding police officer–involved homicides, correlations to local data regarding race, income level and population density may seem obvious. But Brian Finch, professor (research) of sociology and spatial sciences at the USC Dornsife College of Letters, Arts and Sciences, suggests that less obvious factors, such as the level of municipal debt in the city in which the homicide occurred, are worthy of careful examination.

The term "officer–involved homicide" refers to any death involving a police officer, whether intentional, such as an officer shooting a suspect, or unintentional, such as an officer accidentally dropping his gun, which fires and kills someone.

For the past seven years, Finch has supported the development of the Fatal Encounters website, the most comprehensive online database of officer-involved homicides, dating back to 2000. Free and accessible to users, including the public, via maps and spreadsheets, it includes the location of the homicide, cause of death, the poverty level of the area in which the killing took place, and the race and age of the decedent, as well as a brief description of the deadly incident, if available.

Now, Finch is preparing to launch a new repository for police homicide data: the National Officer-Involved Homicide Database (NOIHD). NOIHD is an expansion of the Fatal Encounters website and will provide more data, such as whether the deceased was armed or unarmed, the education and training requirements for the officers in that department and who conducted the autopsies of the deceased. The NOIHD information is collected from police departments, hospital emergency departments, the FBI, census data, state governments and other sources.

"People usually study police homicides in large aggregations, such as by county or state," says Finch. "This expansion of the Fatal Encounters database is adding data at the police department level. That has allowed me to more accurately control for things like whether the level of gun ownership in a state has a relationship to crime at a more disaggregated level, and if so, whether there's also a correlation with police violence."

Crime down, officer homicides up

Finch became involved with Fatal Encounters, which is run by journalist D. Brian Burghart, in 2014 after reading a Washington Post article about the website's tracking of officer homicides that sparked widespread interest in Burghart's work. Finch brought in several researchers from USC Dornsife's Center for Economic and Social Research (CESR) to work on the site's data compilation, and the team began to create NOIHD. Burghart himself also joined CESR as a research associate.

Finch's research has yielded some interesting, albeit sometimes seemingly counterintuitive findings. For example, a recent uptick aside, crime rates in the United States have been steadily and substantially declining for four decades, even though officer-involved homicides have been on the rise over the past 20 years. Perhaps surprisingly, some of the large police departments, like New York's, have not seen an increase in their population-adjusted rate of homicides.

"We're seeing officer-involved homicides become more common in suburban, and even rural areas, rather than cities. And there doesn't seem to be a correlation with crime rates," Finch says.

Location, education and money

Some correlations in officer-involved homicides are rather straightforward, Finch says. For example, departments that require officers to have more training and higher levels of education tend to have lower levels of homicide. Also, deaths often have a racial component; both the decedent's race and the racial makeup and segregation levels of the surrounding area are factors in police killings.

Finch adds that the relationship between law enforcement and other departments also matters. For example, in more than 50 California counties, coroner or medical examiner activities are overseen by the sheriff's office rather than the county government.

"Two coroners in San Joaquin County resigned because they said the sheriff's department pressures them to not report things that are law enforcement-related," Finch says. "We're finding that police homicides in our database are less likely to show up in official sources if the medical examiner or coroner is overseen by the sheriff's department."

Municipal debt may also play a role in police homicides, though Finch says he is still exploring that correlation.

"Someone sent me research that when there's debt, police officers become more proactive in policing poorer areas by constantly imposing traffic fines and doing low-level stops; they're over-policing because they need the money," he says. "And it's become apparent that a lot of these homicides—and violence in general—result from low-level stops that are just completely unnecessary."

Finch notes that there are different types of officer-involved deaths. For example, a car chase that led to a deadly crash might be classified as an accidental death, even if the true culpability has never been established. On the other hand, shootings and asphyxiations are more commonly defined as officer-involved, although many of them never find their way into official government databases.

Finch adds that NOIHD will also include statistics on officers killed in police incidents, but the data will come from the FBI's Law Enforcement Officers Killed and Assaulted database. Because there is a three-year window of time between an incident and when the information surrounding it is released to the public, there will be a lag in data reporting.

A user-friendly version of the database, facilitating simple searches with a focus on local statistics, is expected to be launched in August.