Tamir Rice, a 12-year-old Black child, was playing with a toy pellet gun in a Cleveland park when a police car arrived on the scene. Within moments of exiting his squad car, officer Timothy Loehmann shot and killed Rice. The surveillance video of the November 2014 shooting garnered worldwide attention, and Rice remains a symbol for the Black Lives Matter movement.
As is the case with many high-profile police killings, most after-the-fact reports focused on the incident itself and the officer’s record: Why didn’t Loehmann give any warning before shooting? Would he have done the same to a white child? Why was the officer hired in the first place, given he had been deemed unfit for duty by a different police department?
These are legitimate questions. But it’s possible the most important factor in Rice’s killing was what happened in the moments before the police officer arrived on the scene.
Like the majority of police killings of unarmed civilians, this incident began with a 911 call. The civilian who called 911 on Rice initially reported a Black male with a gun in a park, but then clarified the initial description, saying that Tamir is “probably a juvenile” and that the weapon is “probably fake.” However, according to police records, that clarifying information did not get passed on to responding officers. All the information Loehmann and his partner heard from their dispatcher was, “We have a Code 1” — the department’s highest level of urgency.
That error may have been the difference between life and death for a child.
When Paul Taylor, a former police officer, use-of-force training instructor, and now a criminologist at the University of Colorado Denver, found out that the 911 dispatch information about Rice had been wrong, he decided to run an experiment.
Taylor put 300 police officers representing 18 agencies in two states through an interactive firearms training simulator. All the officers were told about a “possible trespass in progress.” Then some were told that the “subject appears to be holding a gun” and others that the “subject appears to be talking on a cellphone.”
When the officers arrived at the scene, they saw a man matching the description of the suspect with his hands in his jacket pockets. For half the volunteers, the man quickly pulled a cellphone out of his pocket to film the officers; for the other half, the man pulled out a handgun and pointed it at them. The officers had to make a split-second decision to shoot or not shoot, with their virtual lives at stake.
The results were dramatic. Six percent of officers who had been advised that the subject appeared to be talking on a cellphone ended up shooting the man who attempted to film them with his phone. But 62 percent of the officers who were told the suspect had a gun did the same. In other words, officers who were told the man had a cellphone were 10 times less likely to shoot an unarmed suspect than those with incorrect information. (In the scenario where the suspect drew a gun, 100 percent of the officers shot the suspect, regardless of what dispatch told them.)
“What blew me away is that these results held for all officers no matter what,” Taylor told me. “It didn’t matter how much experience you had. It didn’t matter if you were on a SWAT team. Getting the wrong information universally increased the risk of making an error.”
Findings like this one do not excuse police officers of wrongdoing. Nor do they suggest that anti-Black racial bias doesn’t play a huge role in police shootings — it most certainly does. What studies like this (and others) demonstrate is that when it comes to police violence and aggression, the officer-civilian interaction itself is only part of the story.
Of the 50 million Americans who came into contact with the police in 2015, about half were the result of citizen-requested police services, usually through an emergency call number. And 83 of the 153 police killings of unarmed civilians that year began with a 911 call. Research on the 911 system is scarce and imperfect (that’s putting it lightly), so we don’t know for certain how many of these calls contained incorrect information. But the experts I spoke to mentioned data points — like the proportion of calls downgraded by officers once they arrive on the scene — and examples like the killings of Rice, Francisco Serna, and Fridoon Rawshan Nehad and the arrest of Henry Louis Gates Jr. as evidence of the severity of the problem.
“We will never know what would have happened to Tamir Rice if the officer had been given a different image of what was happening,” says Rebecca Neusteter, the executive director of the Health Lab at the University of Chicago’s Urban Labs. “But I would like to believe he would have approached that situation very differently if he was aware this could just be a kid playing in the park.”
The emergency call taker who relayed the incorrect information in the Rice case was temporarily suspended by Cleveland’s police chief for “violating protocol.” But for Neusteter and others who have studied the role of 911 call takers and police dispatchers within the American criminal justice system, the Rice killing isn’t a one-off example of a bad call taker gone rogue — it is the product of systemic flaws in how call takers are trained that amplify the risk officers perceive when they enter a given situation. Addressing those flaws will be essential to the success of any police reform agenda.
Emergency call takers also decide whether police should be sent into a given situation in the first place. Thus, as communities develop alternatives to traditional police response — as many cities are already doing — their role may evolve into that of a public safety quarterback who will be tasked with the all-important role of sending the correct first responders.
Yet call takers are undertrained, underpaid, and underresourced. They are treated as though their role is no different from that of an administrative assistant. And they are ignored in most conversations about policing and criminal justice reform. That’s a shame given the essential role they play in our public safety system.
“911 call takers are gatekeepers not only for police but the entire criminal justice system,” says Neusteter. “We need to start treating them that way. We can’t solve any of our public safety problems without taking care of call takers.”
How the call-taking system can amplify the risk of a dangerous police encounter
To understand an error at the center of Tamir Rice’s killing, you first have to know how the 911 system works.
When you call 911 to summon police, the person you are talking to is generally neither a police officer nor a dispatcher directly in charge of sending police to a scene. Instead, you are talking to a call taker who’s in charge of collecting the relevant information about the incident and the suspect and then classifying the incident according to a list of predefined categories like “suspicious person,” “breaking and entering,” or “active shooter.” That incident type, along with some descriptive information about the suspect, is forwarded to a police dispatcher, who then relays it to responding officers.
This sounds like a fairly innocuous system. But according to Jessica Gillooly, a former call taker and research fellow at the Policing Project at New York University Law School who studies the role of call taking in the criminal justice system, it has a glaring flaw. Call takers are trained and incentivized to think of minimizing potential safety risk to police officers as their highest priority. That means if a caller is uncertain or ambiguous — for instance, simultaneously speculating that the event unfolding could be either a man at a park with a gun (a potential violent threat) or a kid playing with a toy gun (a clearly innocuous act) — call takers are more likely to classify the incident as more serious to ensure officers are prepared for the worst-case scenario.
“There’s a huge training emphasis that essentially tells call takers, ‘You’re safer and better off by sending a police over-response,’” Gillooly tells me. “The big fear is that you don’t send a big enough or serious enough response and something bad happens. There’s no mention of the idea that maybe sending an over-response could also produce a really bad outcome.”
This emphasis produces systemic police over-response. Scholarly research has found that between 20 and 40 percent of all crime calls that 911 call takers enter are downgraded by officers once at the scene. In other words, officers routinely arrive on the scene primed for a far more dangerous, serious encounter than actually exists.
In some cases, this means officers end up killing unarmed civilians like Rice, Serna, and Nehad, each of whom they were led to believe had weapons. More commonly, the result is the sort of humiliation, fear, and aggression that can occur when officers believe they are entering a situation far more serious than it actually is.
“I think about the current state of call taking and dispatching as a game of Telephone,” says Neusteter. “Often, the end result is very different than the original message. And that’s a huge problem. Public safety is too important to leave to a game of Telephone.”
The gatekeepers of our criminal justice system
911 call takers don’t just impact incidents between police officers and civilians; they also determine whether police are sent out in the first place.
Some 240 million calls are made to 911 every year. Tens of millions — possibly hundreds of millions — more are made to non-emergency and alarm lines. In each case, a call taker’s first job is to play the role of gatekeeper: either assign the call to the appropriate first responders or try to resolve the situation on the spot if it does not require immediate assistance.
In the real world, however, this gatekeeper function tends to devolve into a just-send-the-police function. Most jurisdictions have only three types of first responders: fire, medical, and police. And typically there are narrow, predefined criteria for sending in firefighters or EMTs. If those criteria haven’t been met and the situation can’t be easily resolved over the phone, the call taker only has two options: send the police or send no one.
Faced with this choice, call takers will usually opt to send the police for a simple reason: They face severe punishment and liability if they don’t and something bad happens.
“There are situations where if it’s not a clear-cut need for fire or ambulance service, sending law enforcement is the only legitimate response,” says April Heinze, a former call taker and call center director, and current 911 operations director for the National Emergency Number Association (NENA). “That’s not because the call taker wants to send police — they are constrained by local protocol.”
Gillooly, the former 911 call taker and researcher, says she rarely denied police services no matter how benign the situation seemed. She describes a call from someone who found it suspicious that an older Asian man was walking on the side of the road; another about a dispute over a pet peacock defecating on a neighbor’s front lawn; and one from a man who felt uncomfortable at the bus station because a Black teenager’s jeans were hanging too low.
“In most of these cases, sending the police is the only option you really have,” Gillooly tells me. “The informal motto among most call takers is, ‘When in doubt, send them out.’”
Sending police to situations like these can have devastating consequences. That’s why a central plank of the “defund the police” campaign is to reimagine public safety such that police are no longer the default response to all of society’s ills. Instead, activists point to a variety of potential non-police first responders, from trained mediators to crisis specialists to community patrols, that would be better suited to address problems like homelessness, mental illness, and traffic accidents.
In the wake of recent protests against police violence, cities like San Francisco, Oakland, Portland, Denver, Minneapolis, Albuquerque, and Los Angeles are developing their own civilian first responder programs. And Sens. Ron Wyden (D-OR) and Catherine Cortez Masto (D-NV) recently introduced the CAHOOTS Act — named after the much-applauded initiative in Eugene, Oregon, that sends unarmed crisis specialists instead of police to address noncriminal 911 calls — that would provide federal government support for such programs.
But even if these alternative programs are successful politically, they will only succeed logistically if 911 call takers can clearly distinguish between incidents that require sending in police and those that don’t.
“We expect our call takers to make really important judgment calls,” says Steve Zeedyk, a call center supervisor in Eugene who works closely with Cahoots. “There are many jurisdictions where if someone calls and wants an officer, they get an officer. Our call takers screen at a much higher level to determine whether police really are the right response. That’s why we’re able to make good decisions deploying the resources we have.”
My conversations with Zeedyk and others made clear that 911 call takers will be crucial to the success of any non-police response efforts. “The 911 system needs to be part of the conversation as cities think about how to set up alternate public safety initiatives,” says Ayesha Delany-Brumsey, director of the Behavioral Health Division at the Council of State Governments Justice Center. “Call takers are going to make consequential decisions about what responders get called in where.”
How to fix our 911 system
A few modest changes and investments could go a long way toward addressing the 911 system’s tendencies to default to police and amplify the risk of police over-response.
As Gillooly points out in a recent paper on the subject, the technology that call takers use to transfer call information could be redesigned to include fields that capture a situation’s level of ambiguity and uncertainty, signaling quickly to the dispatcher and police that the information they’ve been given may be wrong. Training for call takers could be more comprehensive and include a greater emphasis on asking clarifying questions — like “are you sure that the gun is real?” — before classifying an incident. And incentives for call takers more broadly could be changed to incorporate the social costs of sending a police over-response.
A more sweeping solution would be to invest in significantly upgrading the technology that 911 call takers use. Imagine how many problems with the current 911 system would be mitigated if call takers could receive pictures or videos of a given situation as it is happening and then forward them directly to the responding officer — or use them to determine that police aren’t needed for the situation at all. That’s part of the vision behind “Next Generation 911,” an initiative spearheaded by the US Department of Transportation’s National 911 Program to upgrade the emergency call system nationwide.
According to experts at NENA, the biggest obstacle to Next Gen 911 deployment is inadequate funding. Before Covid-19, about half of jurisdictions in the US were slated to have Next Gen core services by the end of 2020 and 85 percent by 2025; however, the pandemic’s impacts on state and local government budgets may create a shortfall of funding and thus delay deployment.
A modest federal investment could change that. According to the National 911 program, the cost of national deployment of Next Gen 911 comes out to about $12 billion over five to 10 years, a relatively small drop in the bucket of the federal budget.
“It’s about time to move 911 technology into the 21st century,” says Brian Fontes, CEO of NENA. “With so many 911 calls originating from smartphones, there is so much potential information we could gather that is essential to responding to an emergency.”
Call centers could also implement “criteria-based dispatching,” a script-based set of questions that guide the call-taking process. This would mean that both the level of police response and whether police are sent in at all would be left up to predefined criteria instead of the subjective discretion of the call taker, which could be subject to all kinds of momentary biases. The criteria-based dispatching model is often used in medical and firefighting dispatching centers and has been credited with curbing over-response. The approach is being piloted for policing in a handful of cities including Seattle, Tucson, Houston, and Washington, DC.
With criteria-based dispatching, the important consideration is to draw the criteria boundaries such that police forces aren’t the default response. For instance, in Seattle, part of the dispatch criteria makes a strict distinction between “suspicious activities” and “suspicious persons”; if the caller can’t definitively name a specific suspect behavior that a given person is engaging in, the call taker does not dispatch police.
Some places have gone a step further. Houston 911 call-taking scripts involve mandatory questions to assess whether the given incident involves someone experiencing a mental health crisis. If a case does involve a mental health component, it is flagged for dispatchers. And for those cases, the city employs a handful of mental health clinicians to sit with dispatchers and help them determine the appropriate first response: a civilian clinician team, a co-response team of police and clinicians, or a police team.
The result is that of the 40,000-plus calls that were flagged by call takers as having a mental health component in 2019, only 0.5 percent ended in an arrest, according to Wendy Baimbridge, assistant chief of the Houston police’s mental health division. That’s partly because Houston sometimes sends non-police first responders, but it’s also because when police officers do enter such situations, they are fully aware that what they are dealing with is probably a mental health crisis.
“Call takers can’t possibly train for everything,” says Baimbridge. “And they certainly don’t have the time to do a full mental health assessment. That’s why we need mental health clinicians on the floor to play that role.”
Of course, without the availability of non-police first responders, reforms like these will only go so far. For the many situations that require some kind of trained response, call takers can’t do much except call the police unless they have alternatives available.
In addition to these specific reforms, the experts I spoke with called for a cultural shift in how we as a society view, compensate, and treat emergency call takers and dispatchers. Only 20 states have even minimum training requirements for call takers and dispatchers, and even fewer provide funding for that training. In most states, call takers make less than $50,000 per year with scant benefits. And they routinely experience burnout, high stress levels, and PTSD from their work.
“The 911 system has been completely undervalued, underfunded, and underresourced for 50 years,” says Neusteter. “The technology is terrible. The training, benefits, and occupational standards are subpar. Call takers have not been set up for success institutionally.”
That’s a shame because call takers are the first point of contact, the most common reference point, and the gatekeeper for our entire criminal justice system. As cities and communities across the country wrestle with how to change policing, it’s more important than ever they invest in 911 call centers that are better equipped, better trained, and better suited to handle the range of responsibilities they will be tasked with.
“Over the years, 911 has been treated as a stepchild of the public safety community,” says Fontes. “That needs to change.”
Will you become our 20,000th supporter? When the economy took a downturn in the spring and we started asking readers for financial contributions, we weren’t sure how it would go. Today, we’re humbled to say that nearly 20,000 people have chipped in. The reason is both lovely and surprising: Readers told us that they contribute both because they value explanation and because they value that other people can access it, too. We have always believed that explanatory journalism is vital for a functioning democracy. That’s never been more important than today, during a public health crisis, racial justice protests, a recession, and a presidential election. But our distinctive explanatory journalism is expensive, and advertising alone won’t let us keep creating it at the quality and volume this moment requires. Your financial contribution will not constitute a donation, but it will help keep Vox free for all. Contribute today from as little as $3.