High-profile police shootings in cities have set off protests across the country. But police in rural areas have fatally shot about 200 people a year since 2015, our investigation found, with little scrutiny. Rural shootings are in many ways similar to urban shootings: the dead are mostly men who are often suffering from drug addiction or mental health problems. But in rural areas, those shot by police are mostly White. The agency with the largest number of rural shootings was the Kentucky State Police; its troopers don’t wear body cameras and have never been prosecuted for a shooting.
Our investigation – which was co-reported with the Kentucky Center for Investigative Reporting and co-published The New York Times and the Lexington Herald-Leader – found that police shootings in isolated areas are rarely captured on video, and many rural officers don’t wear body cameras. We found many disputes about what happened during shooting incidents. Shortly after our stories were published, Kentucky Gov. Andy Beshear requested millions of dollars to equip troopers with body cameras. Some local media used our reporting on the troopers to inform their own work and both editorial boards and advocacy organizations including the ACLU called for changes at the state police. Having The New York Times as a media partner and two publishing partners in Kentucky ensured our reporting would be widely seen, both locally and nationally.
GIS tools: The analysis made extensive use of GeoPandas for the initial analysis and PostGIS for the final analysis and review.
Geospatial joining of disparate datasets: We combined geographic data from the US Census, police shooting data from Kentucky as well national datasets published by the Washington Post and Mapping Police Violence, and applied an urban/suburban/rural classification scheme published by the US Department of Housing and Urban Development.
The project also involved significant work with records requests and document-based research, as well as legal challenges to enforce public-records laws.
What was the hardest part of this project?
The data reporting for this project was laborious. Reporters Alysia Santo and Weihua Li used a database to mark the locations of police shootings between 2015 and 2020. They then worked to determine whether the shootings had taken place in rural, suburban or urban areas. There is no universal classification for what qualifies as a rural area, so they tested the police shooting data in several different government systems. They settled on the Department of Housing and Urban Development’s system, which uses survey information from people about their communities to define rural areas. In Kentucky, figuring out who had been fatally shot by officers, and what exactly happened in each incident was no small feat. State police wouldn’t provide answers, so The Marshall Project and KyCIR used public records requests to build a database of fatal shootings by troopers. Obtaining the documents necessary for this required a legal battle, including multiple appeals. Eventually, the Kentucky attorney general found that the agency violated the state’s open records law when it denied The Marshall Project’s request. Santo submitted records requests for each of the 41 fatal shootings by Kentucky state troopers in the previous five years. She built a database with those records, which allowed the reporters to search what kind of call officers responded to and whether the people who were shot had been armed, had a history of mental illness or were intoxicated. Of the 41 shootings, 33 were in rural areas of Kentucky. Despite restrictions imposed by the COVID-19 pandemic, Santo and R.G. Dunlop, a reporter with the Kentucky Center for Investigative Reporting, interviewed more than 100 people and reviewed dozens of court cases and thousands of pages of police investigative reports. Santo researched obituaries, and used Facebook and other public records databases to find family members of the
What can others learn from this project?
We hope this project will inspire other journalists to look at police shooting deaths in their states that aren’t getting attention and figure out what happened. They could replicate the reporting we did in Kentucky in their own states by documenting all police shootings and using the Housing and Urban Development system to define which ones were in rural areas.