Skip to main content

Databases Vastly Underreport Police Homicides

But a CDC database that began in 2003 may help solve that problem.
(Photo: Kevork Djansezian/Getty Images)

(Photo: Kevork Djansezian/Getty Images)

Police homicides are drastically underreported in the most-used statistical databases. Fortunately, there exists a better alternative, public-health researchers report today in the American Journal of Public Health—a national system that's been in place since 2003, and that, as of last year, 32 states participate in.

"Recent media attention to the killings of civilians by police has given rise to a national conversation on the use of lethal force by law enforcement: when it is justified and how best to reduce its use while ensuring the greatest possible safety to police," writes a team of public-health researchers led by Harvard University School of Public Health's Catherine Barber.

Unfortunately, two main sources of data on police killings—the Federal Bureau of Investigation's Supplementary Homicide Reports, part of its voluntary Uniform Crime Reporting system, and the Centers for Disease Control and Prevention's (CDC) Vital Statistics database—grossly underreport the numbers, sometimes because reports simply fail to mention police involvement. For researchers, that makes it difficult to understand who police are killing, why it's happening, and what could be done to reduce deaths.

On average, police killed 0.24 people per 100,000 every year.

On the other hand, a CDC initiative that began with just six states in 2003—called the National Violent Death Reporting System—offers some hope. NVDRS draws on both police reports and a death certificate or coroner's report for each violent death, and it includes two brief narratives based on those reports, along with a simple "type of death" indicator—homicide, suicide, or accident, for example. But how good is that indicator?

To evaluate NVDRS alongside the others, the researchers needed an independent way to identify whether a violent death was a police homicide—so they sat down and started reading. Using a set of three criteria, including whether a killing occurred in the line of duty, the team identified 1,552 police-justified killings. Of those, NVDRS pinpointed 92 percent "legal intervention homicides," another, somewhat more precise term for justified killings. Meanwhile, only 48 percent of killings the researchers found were listed as police homicides in the Supplementary Homicide Reports data, and only 58 percent showed up in the CDC's Vital Statistics file as police killings.

The vast majority of the cases the team identified as police homicides, but that NVDRS did not, were in four states: Colorado, Georgia, Kentucky, and South Carolina. On average, police killed 0.24 people per 100,000 every year, though at 0.48 per 100,000 per year, the number for African Americans was much higher.

"The NVDRS captured more than twice the legal intervention deaths reported by SHRs and 71% more than reported by Vital Statistics," Barber and her colleagues write. "Expanding the system from its current (as of 2016) 32 states to 50 and the District of Columbia will be an efficient way to supply the nation with both an accurate count of police homicides and detailed information from the coded and narrative data on the people, weapons, and circumstances involved."


Quick Studies is an award-winning series that sheds light on new research and discoveries that change the way we look at the world.