Police Kill Twice As Many People As We Thought

By Tim Dickinson, Rolling Stone

More than half of all police killings since 1980 do not appear in official government data, according to an explosive new study in The Lancet, a top medical journal. The researchers reveal how “systemic misclassification” in the federal database that tracks the causes of death in America has produced, over four decades, an undercount of more than 17,000 deaths at the hands of police. The proportion of undercounted police killings of Black Americans is even more extreme, the research shows, rising to 60 percent.

The Lancet study casts American police, unequivocally, as a threat to public health. The risk of death-by-cop for an American man in 2019, according to the paper, was higher than the risk of death by testicular cancer, appendicitis, or sexually transmitted disease. These dangers weigh disproportionately on the Black community, as the study emphasizes: “The police have disproportionately killed Black people at a rate of 3.5 times higher than white people.”

The research for The Lancet paper was conducted by the Institute for Health Metrics and Evaluation, which is housed at the University of Washington Medical School, and which catalogs global disease and injury, studying more than 200 leading causes of death.

In an exclusive interview with Rolling Stone, IMHE researchers said they were motivated to resolve “inconsistencies” in public counts of deadly encounters with police. Independent groups like Mapping Police Violence — which rely on open-source data including news reports of police shootings — have long tallied higher numbers of killings by police than those recorded in the federally funded National Vital Statistics System. NVSS includes an exhaustive database of American causes of death, compiled from death certificates filled out by doctors, medical examiners and coroners. It has tracked deaths from law enforcement since 1949.

“Last year, after the murder of George Floyd, we reviewed our police-violence data, and we found other literature that supported that NVSS was undercounting,” says Eve Wool, a co-lead author of the study. “And we went and sought other sources to correct it.” The IHME researchers selected high-quality open-source datasets on police violence, including Fatal Encounters and The Counted, and compared those studies to the federal record, building a statistical model of killings by police over time. This effort yielded — as The Lancet touts in an accompanying editorial —  “the most accurate and comprehensive assessment of deaths attributable to police violence in the USA to date.”