The systemic problems within police departments across the nation have been a hotly debated topic for several years now. After the death of Michael Brown in Ferguson, Missouri, kicked off protests that would become the foundation of Black Lives Matter, public attention to racial profiling and other policing tactics picked up considerably. For those who want reform in America's police force, the findings of a recent study will probably serve as further evidence that the system needs to change. In a study released this week, researchers at Harvard found that police-related deaths are woefully underreported.
The study's authors came to their conclusion by examining reports issued by the National Vital Statistics System (NVSS) and comparing those numbers to media collected by The Counted, an online database run by The Guardian that compiles news stories of police-related deaths.
When the Harvard team looked into the reason for this discrepancy, they found the disparity came from coroners and medical examiners misreporting on hundreds of death certificates the cause of death as "assault" rather than "legal intervention," the term used by the NVSS to designate a police-related fatality.
Notably, a cause of death labeled "legal intervention" does not entail any implication of wrongdoing or unlawful use of force. It's the term used in all deaths that occur as a result of a police officer's actions, regardless of the details surrounding any particular altercation. Still, in hundreds of cases, coroners and medical examiners listed "assault" rather than "legal intervention" as the cause of death.
The study concluded that the real number of deaths that should have been labeled "legal intervention" was 1,086, far higher than the 530 reported by the NVSS. Examining death certificates, the research team found 599 of the deaths tracked by The Counted had been mislabeled, which explains the substantial gap between the two numbers.
It wasn't the only jarring finding for the Harvard team. Doctoral candidate Justin Feldman, one of the study's authors, tells Bustle:
It was also surprising that some states did a much worse job counting killings by police, while others did better. In Oklahoma, where more than 30 people were killed by police in 2015, none of these deaths were counted. But other states such as Washington properly counted a large majority of these deaths.
In other words, according to the NVVS, Oklahoma would appear to have a 2015 record of no police killings. As Feldman points out, that's an egregious factual error.
Feldman also tells Bustle that in order to change this flawed system, "there will need to be a few different measures taken." One of those ideas is to replicate government regulations that already exist for other major health concerns, such as mandatory reporting of the Zika virus and lead poisoning. "Killings by police can be added to these lists of notifiable conditions," says Feldman.
The authors also found the likelihood of a misreported cause of death was higher in lower-income counties. Their findings did not conclude that race or ethnicity were necessarily correlated to a greater chance of mislabeled death certificates. They also found no statistical evidence that medical examiners were more likely than coroners to misreport cause of death, or vice versa.
Since NVSS data relies on designations made by medical examiners and coroners, the possibility of corruption in at least some of these cases seems a possibility. The Harvard study does note that "journalists have revealed multiple incidents in which law enforcement agencies failed to release pertinent documents to death investigators for in-custody deaths or pressured death investigators to make a finding of non-homicide," though they also add it is impossible to know how frequently such events occur. Feldman tells Bustle, "Our study was not able to draw conclusions about why coroners and medical examiners misreport deaths."
One thing that is certain: With such enormous gaps in their data, the NVSS needs to change how it tracks and reports police-related deaths.