Police officers are killing more Americans every year than official statistics indicate, new research out Thursday concludes. The study compared government data to other records maintained by outside organizations and found a wide gap, suggesting that more than half of police killings have been unreported by the federal government in the last 40 years.
Criminal justice organizations and journalists have long criticized the U.S. and local police departments for a lack of transparency around police-related deaths—an issue that has only gained more attention in the wake of high-profile deaths involving Eric Garner, Breonna Taylor, and George Floyd, among many others. To date, there is no specific federal database keeping track of these killings nor why they may have happened. But several media outlets and independent groups have formed their own record-keeping operations, reliant on openly available information like news reports and public record requests.
This new research was led by scientists at the University of Washington School of Medicine, as part of a long-running project that tries to tally the worldwide toll of major health conditions and causes of deaths, the Global Burden of Disease Study.
The companion study, published in the Lancet on Thursday, looked at data from the National Vital Statistics System (NVSS), which provides the most complete information on birth and death records collected and maintained by the federal government. These records do include deaths that are officially reported as due to police violence by medical examiners or coroners. Then the team compared this data to the data collected from three groups keeping track of police-related deaths, Fatal Encounters, Mapping Police Violence, and The Counted.
The oldest of these watchdog projects, Fatal Encounters, only dates back to 2000. But the researchers still found a significant disparity between the official and unofficial numbers when they could be directly compared. Using this disparity as a baseline for their model, they estimated that from 1980 to 2018, 55.5% of all deaths involving police violence had not been counted officially as such. In raw numbers, that would amount to an estimated missing 17,100 deaths out of 30,800 total deaths. Moreover, they estimated that there had been another 1,190 deaths that occurred in 2019.
“Recent high-profile police killings of Black people have drawn worldwide attention to this urgent public health crisis, but the magnitude of this problem can’t be fully understood without reliable data. Inaccurately reporting or misclassifying these deaths further obscures the larger issue of systemic racism that is embedded in many US institutions, including law enforcement,” said co-lead author Fablina Sharara of the Institute for Health Metrics and Evaluation, which maintains the GBD project, in a release from the university.
Indeed, as other research has shown, Black Americans were more likely to experience fatal police violence than other groups in this study (3.5 times more likely than white Americans); they were also slightly more likely to have their deaths uncounted as police-related. But there was less disparity between the official and unofficial count for Americans of other ethnicities. Men were also much more likely to be killed by police than women (30,600 estimated deaths in men and 1,420 deaths in women from 1980 to 2019).
Critics argue that there remains a frustrating lack of information on these violent encounters from law enforcement, even with the advent of body cameras. And while police are not killing the general public en masse, some research has estimated that police violence is the sixth leading cause of death among young men in their late 20s. Notably, the study found that the rate of police killings had actually climbed from the 1980s up through the 2010s, despite crime in general substantially falling during the same period. Many other peer countries report far less police violence against civilians.
The GBD project has historically relied on NVSS data for their estimates of police killings in the U.S.. But the authors say they had already switched to their current method for the latest report on 2019, which was published in October 2020, and will use it moving forward. In 2019, they estimated there had been 293,000 deaths caused by police worldwide. And though there are other sources of official state-level data that may provide a more accurate picture of fatal police violence than the NVSS, they say that the government’s numbers can’t really be trusted for the time being.
“Our conclusions from doing this research and from other studies before ours is that until changes are made to improve accuracy for this cause in the vital registration system, those looking for accurate data on police violence should utilize open source databases,” senior study author Mohsen Naghavi, director of Subnational Burden of Disease Estimation at IHME, told Gizmodo.
This article has been updated with comments from study author Mohsen Naghavi.