ON MARCH 31ST 1931 a plane carrying Knute Rockne, a legendary American football coach at the University of Notre Dame, crashed into a wheat field outside Bazaar, Kansas, killing everyone on board. The incident sparked a national outcry and inspired a new standard for investigating aviation accidents. A government agency, which eventually became the National Transportation Safety Board, or NTSB, assumed responsibility for gathering relevant data and establishing best practices.
America has not responded to the continued killing of black Americans by police—Rayshard Brooks, George Floyd, Breonna Taylor, and countless others before them—with the same urgency. As a result, officials trying to reform law enforcement face a dearth of reliable data, such as on racial bias. Without it, America will end up, at best, where it is now: in a policing debate dominated by ideology not information.
In 2017 the NTSB received $106m and deaths from commercial air travel in America amounted to exactly zero. That same year, total funding for research on policing stood at just $11m, while police killed roughly 1,000 Americans. It is long past time to invest in better data to inform decisions about policing in America.
Make no mistake, policymakers are right to forge ahead with reforms even with imperfect metrics. America needs a wholesale reimagining of public safety—and massive, targeted investments in social programmes—to even begin repaying the debt owed to black communities after generations of discrimination and dispossession. More equitable policing is just one piece of the equation.
Already, the City Council in Minneapolis voted to disband the police department that employed George Floyd’s murderer. San Francisco’s mayor, London Breed, plans to replace a portion of her city’s police with unarmed professionals trained to respond to non-criminal matters such as mental-health crises and disputes between neighbours.
We do not know exactly what these rebuilt systems will look like, although focusing on more tailored, less violent solutions to social ills such as homelessness or substance abuse is a good place to start. There will undoubtedly be setbacks and successes along the way. Accurate information produces better decisions and reduces the risk that policies will be ineffective.
Quality data to measure the effects of policy changes can identify successes to be scaled up. Yet if some reforms falter and we lack the data to correct course, public sentiment may turn against change as quickly as it shifted in favour. The backlash against ineffective reforms may smother a potentially transformative movement.
The structure of American policing does not lend itself to the standardised data collection needed for rigorous analysis. The United States has roughly 18,000 law-enforcement agencies, each with a unique approach to collecting, storing and aggregating its data. Some agencies barely collect any data at all. As a result, we lack answers to basic questions such as how many police interactions occur each year or the number of incidents that send someone to the hospital.
This allows deeply ineffective, even counterproductive practices to continue unchallenged. In 2006 Malcolm Gladwell wrote about a man dubbed “Million Dollar Murray” in Reno, Nevada. For decades, the city responded to Murray Barr’s chronic homelessness and alcoholism with the only tools available: policing and hospitalisation.
After his death, Reno police estimated that Murray had cost taxpayers over $1m in ineffective interventions. As America undertakes a monumental reinvention of public safety, policymakers need quality data to avoid these kinds of mistakes. Knowing how we use or misuse various public-safety tools will also help us avoid disastrous pitfalls.
One of those pitfalls is failing to recognise when new policies produce unintended harms. For example, by 2016 some 23 states had passed laws to prevent employers from screening people with criminal records. The idea was to increase job opportunities for people recently out of jail, which would give them an income and reduce recidivism.
But because of the racist assumption that black applicants were probably criminals if they could not show they had not gone to jail, the policies actually led to increased racial discrimination in hiring. America’s states, the “laboratories of democracy”, are about to start unprecedented experiments in policing, on a much greater scale, without the ability to tell if they are working.
Effective policies can also fail when we lack enough data, or the right data, to appreciate their successes. Consider a lesson from history. Following the Civil War, the federal government began a programme called Reconstruction: a massive effort to rebuild the defeated Confederacy and enforce the rights of newly-freed slaves. For a while, it worked. South Carolina, a majority-black state, elected a majority-black legislature. Land ownership and literacy rocketed.
But faced with mounting costs and flagging political will, few cared to account for the benefits accruing to black people. Politicians such as Daniel Henry Chamberlain, the white Republican governor of South Carolina during Reconstruction, became disillusioned. Writing in The Atlantic, he later claimed that the era had not led to “a single public improvement of any sort.” Federal troops withdrew. Over the next 75 years, without their protection, mobs across the South lynched more than 6,500 black people.
Of course, Reconstruction didn’t fail because of a lack of data. Rather, a backroom deal to stop federal “interference” in the South in return for those states’ support for Rutherford B. Hayes as president in 1877 spelled the end of Reconstruction. But it does illustrate how rapidly political will can abandon black Americans, especially when change requires a sustained commitment. When it comes to supercharged ideological issues like racism and policing, success that cannot be seen is the same as failure.
To stand any chance, champions of change need the fullest possible information to make the case for continued efforts. Experience shows how sound data collection can bring sanity to stubbornly ideological policy debates, such as one on the decriminalisation of cannabis. Policymakers can expect these benefits to apply to policing, as well.
National investment in data collection should go hand-in-hand with analyses that help answer questions about bias, effectiveness and consequences. Individual police departments should require officers to record information each time they stop someone. That should include the perceived race of the person, what led to the stop, any use of force and enforcement outcomes. State legislatures should mandate a baseline level of collection on every non-consensual stop by an officer and, importantly, standardise it in a way that allows for national comparisons and tracking over time, along with the ability to draw causal inferences.
The notion that we need to measure outcomes we care about seems obvious. It was what motivated the sweeping reforms to aviation in the 1930s that enabled the air-travel industry to mature and set the global standard for safety. Yet the same goal eludes us when it comes to policing. America is embarking on a treacherous journey to a reconstructed, more equitable system of public safety. To arrive safely, we need the information that helps us avoid dangerous pitfalls, allows us to correct course and bolsters a sustained case for the change we desperately need.■
Phillip Atiba Goff is the cofounder and chief executive officer of the Centre for Policing Equity and a professor at the John Jay College of Criminal Justice.