Police Killings Happen Far More Often Than What’s Widely Reported. Here’s Why the Numbers Are Off by So Much

A new study of over 30 years worth of medical data reveals the number of deaths at the hands of police officers in the U.S. is more than double the amount that’s been widely reported.

The Institute for Health Metrics and Evaluation (IHME) published the study on Thursday in the medical journal The Lancet, having compared data from the U.S. National Vital Statistics System (NVSS), which provides data on births and deaths in the U.S., to three open-source, non-governmental databases on fatal police shootings.

According to the study, between 1980 and 2018 there were about 30,000 deaths from police violence across the country. On average, that’s the equivalent of 2 people in the U.S. being killed by police officers every single day. (The study does not take into account non-fatal injuries suffered at the hands of police.) But more than half of the total deaths—55%, or 17,100 deaths—were not reported by NVSS.

“It really is obscuring the true issue, which is the magnitude of police violence and the disparities of police violence by race and ethnicity,” Eve Wool, the co-author of the study tells TIME. “We really want to emphasize that the main problem here is the true rate of police violence is [being underreported].”

Researchers also found that Black Americans were as much as 3.5 times more likely to die from police violence than white Americans. Though other studies in the past have found similar rates of under-reporting and racial disparities, the IMHE’s report is the most expansive with regards to the timeframe covered.

“Police violence is a public health problem, the same as racism and to fix a public health problem we need very good information. If we do not have correct information all of our [path to address the issue] will go to the outside of the problem,” Mohsen Naghavi, another co-author of the study says.

The study’s publication comes just a week after the bipartisan talks on police reform fell through; the George Floyd Justice in Policing Act had been stalled in the Senate since its successful passage in the House of Representatives this past March. The bill had called for a ban on racial and religious profiling by police officers, overhaul qualified immunity for officers and set up a national database for police misconduct, along with other reform efforts.

Read more: How George Floyd’s Killing Highlighted America’s Police Reform Failures

“Senate Republicans rejected enacting modest reforms, which even the previous president had supported, while refusing to take action on key issues that many in law enforcement were willing to address,” President Joe Biden said in a Sept. 22 press statement about the failed negotiations.

“The United States has amongst the most, if not the most, killings at the hands of police—[more] than any other wealthy country in the world,” explains Edwin Lindo, a professor and assistant dean at the University of Washington. “What happened to George Floyd is a prime example of why [this study] is so important,” Lindo adds, referring to the first press release the Minneapolis Police Department shared after Floyd’s death, which said he died from a medical incident he was experiencing while in police custody.

Since many medical examiners and coroners work for their respective police departments, the study argues that a level of implicit bias is often at play in the classification of a police death. There’s also an issue with the paperwork itself, Wool says: if someone’s death is to be attributed to a police officer, the information must be written on their death certificate in the certificate’s free text field. Wool and other researchers on the study say that the instructions for inputting that detail are not always clear or consistent across different state and city agencies. This often leads to police killings being misclassified as homicides.

The study’s authors are recommending better training and guidance for medical examiners and coroners who are tasked with classifying these deaths, as well as calling for the use of more open-source data when researching and tallying the numbers. But the goal of the study is also to have a larger impact: “This is not a call for more research. We hope that this information can provide states with accurate complete information that can drive policy change,” Wool says.