Following the murder of George Floyd, The Associated Press issued guidance that reporters should not describe instances where police shot someone with neutral language. It didn’t work.
This story was published in partnership with The Huffington Post.
On Dec. 27, 2021, the Los Angeles Police Department released an antiseptic statement titled “North Hollywood Officer Involved Shooting Incident in North Hollywood.” It said that four days earlier, North Hollywood officers received a report of an “Assault with a Deadly Weapon in Progress” at a business and, when they arrived, shot and killed the suspect and a bystander.
The “officer-involved” language obscured a horrific truth about the police’s actions on Dec. 23. That day, officers responded to reports of someone acting erratically in a North Hollywood Burlington store who was allegedly attacking customers. Officer William Jones Jr. charged into the scene with a high-powered rifle drawn. Despite being told to “hold up” by his colleagues, Jones fatally shot the suspect, Daniel Elena-Lopez, 24. One of Jones’ shots also pierced a wall, killing 14-year-old Valentina Orellana-Peralta, who was huddled with her mother in a dressing room. Jones has been placed on administrative leave. The California Attorney General’s Office will be investigating and will independently review the shooting. Independent counsel for Jones’ police union said he did not believe “hold up” meant that Jones was being told to stop by his fellow officers.
A Dec. 30 LAPD update to its Dec. 27 statement described Jones as “the officer involved in the below incident.” Jones, of course, was far more than neutrally “involved” in an incident that claimed two lives. But several news outlets still adopted the LAPD’s self-interested language. “Per California law, the California attorney general will be investigating and independently reviewing the officer-involved shooting,” wrote ABC News. CNN included a similar sentence in its initial report.
On Jan. 6, Los Angeles Magazine tweeted that Orellana-Peralta was “killed by a stray bullet in a police shootout,” which was both passive language and inaccurate. The only shots were fired by the police: Elena-Lopez was holding a bicycle lock when he was killed by the LAPD. (The magazine later deleted its tweet).
In theory, the media should have left this specific “officer-involved” terminology behind after the killing of George Floyd in May 2020, which spurred perhaps the largest movement in the history of the United States and also prompted newsrooms to reexamine longstanding practices regarding their coverage of communities of color, particularly how they report on police shootings.
Soon after Floyd’s murder, Wesley Lowery, a correspondent with “60 Minutes+” and a former Pulitzer Prize-winning national correspondent for The Washington Post, urged a “reckoning over objectivity,” and cited the phrase “officer-involved shooting” as a prime example of what he called “neutral objectivity.”
“Neutral objectivity insists we use clunky euphemisms like ‘officer-involved shooting,’” Lowery wrote in a June 2020 New York Times op-ed, “Moral clarity, and a faithful adherence to grammar and syntax, would demand we use words that most precisely mean the thing we’re trying to communicate: ‘the police shot someone.’”
On Aug. 24, 2020—one day after a white Kenosha Wisconsin police officer shot and seriously injured Jacob Blake, a 29 year-old Black man—the Associated Press Stylebook instructed reporters to “avoid this vague jargon for shootings and other cases involving police. Be specific about what happened. If police use the term, ask: How was the officer or officers involved? Who did the shooting? If the information is not available or not provided, spell that out.”
The response to Floyd’s murder and growing criticism around the media’s coverage of police violence drove the AP to issue the guidance: “A day after the Aug. 23 police shooting of Jacob Blake, the Stylebook team agreed unanimously to create a separate entry. We published it in Stylebook Online that day, and tweeted it the next day,” Stylebook editor Paula Froke said.
The AP’s tweet detailing its guidance about “officer-involved” was retweeted more than 23,000 times and widely embraced by reporters. “I would love to know who is leading the revolution over at the AP,” tweeted Time senior correspondent Janell Ross. “It can’t be easy.” Indeed, the decision from Stylebook — a guide created by U.S. journalists working for or connected to The Associated Press, described recently by Washingtonian as “a journalism bible”—seemed to signal the end of the media’s decades-old habit of using the phrase.
But we, in collaboration with HuffPost, decided to examine how often reporters have used the phrase since then — as well as the decades leading up to its guidance. To do so, we analyzed approximately 136,000 newspaper articles that ran between 2000 and 2021 categorized under “Death and Injuries By Police” in the LexisNexis database.
We counted the number of times “officer-involved” and the similar “police-involved” appeared in articles. The articles came from 35 major American newspapers, chosen to represent the largest outlets — the AP, The New York Times and USA Today among them — as well as local daily papers across the country. Six of 35 newspapers appeared to have some years of missing data from LexisNexis, but excluding them from our analysis did not significantly change results so we did not eliminate them.
We chose the time period of 2000-2021 because, as the Columbia Journalism Review reported in 2020, “officer-involved shooting” appeared much more frequently starting in the early 2000s. These 20-plus years also allowed us to observe a baseline of usage in the years prior to the emergence of the Black Lives Matter movement in the early to mid-2010s that significantly changed the public’s perception of police violence as well as the media’s coverage of it.
The results were disheartening. Usage of “officer-involved” rose steadily throughout the 2000s. We found an average of 22 uses of “officer-involved” per month, with the phrase appearing at least once in 5% of all articles in 2000. By 2010, there was an average of 34 uses per month with the phrase appearing once in 11% of all articles about police violence.
Into the early 2010s, usage increased more rapidly and remained high later in the decade. In 2013, 15% of all articles about police violence surveyed used “officer-involved” at least once.
Usage of the phrase declined in 2020, but following Floyd’s murder — and, two months later, the new AP guidance — it began to climb back up in 2021. In August of last year, “officer-involved” appeared 56 times, and was used at least once in 8% of all articles about police violence in that month. This represents a return to the rates in the mid-2000s, after usage of “officer-involved” had begun to significantly increase.
We also analyzed active-voice descriptions of police violence, which produce better, clearer reporting. (The AP described “officer-involved” as “vague.”) We counted the number of—and percentage of—uses of “police shot” and “police killed.” The use of either active phrase is consistently significantly lower than the use of passive phrases. In articles about police violence published each month during the 2000-2021 span surveyed, active phrases appeared 100 times or more during just three months.
Consistent use of passive language and the phrase “officer-involved” demonstrates that the media, like many institutions, often does not remain vigilant when it comes to holding police accountable.
According to the Columbia Journalism Review, “officer-involved shooting” first appeared in newspapers in the early 1970s, likely because the Los Angeles Police Department adopted it and police “quickly institutionalized the phrase.” Reporters took their cues from the police — who routinely use “officer-involved shooting” and the acronym “OIS” — and the phrase was adopted by the newspapers where they worked, too.
Decades later, the phrase came under scrutiny from activists and reporters during the headiest days of the Black Lives Matter era from 2014 to 2016, which saw the police killings of Michael Brown, Eric Garner, Philando Castile, Freddie Gray, Alton Sterling and thousands of others. In a July 2014 op-ed, Washington Post opinion columnist Radley Balko wrote that “officer-involved shooting” is “a way of describing a shooting without assigning responsibility.”
Froke told HuffPost that the AP’s internal discussions about “officer-involved” began in 2015, first resulting in its inclusion in the Stylebook’s “Cliches, Jargon” entry—what Froke described as “an extended entry on jargon and cliches, with ‘officer-involved shooting’ as one of many examples to avoid”— in 2017. And then, in 2020, its clearest guidance: Stop using the phrase.
“We believe our entry speaks for itself: Avoid the term. Be specific about what happened. ‘Avoid’ pretty much means ‘don’t use,’” Froke said.
Our analysis showed uneven results after that guidance, after years of waxing and waning usage of the term.
Froke said the AP’s guidance still allows for usage of “officer-involved” in a headline when “space is an issue and editors after trying diligently can’t come up with a strong alternative,” so we excluded usage of the phrase in headlines in our analysis.
The media is not a neutral participant in covering police violence—what is or is not covered is a choice based on subjective decisions by reporters and editors. The number of police shootings has remained consistent over the past several years, even as usage of “officer-involved” increased. This meant that increases—or decreases—in the phrase’s usage couldn’t be driven by major swings in the number of people killed by police. Indeed, when The Washington Post updated its “Fatal Force” database recently, it noted “police nationwide have shot and killed almost the same number of people annually — nearly 1,000 — since the Post began its project” in 2015.
We separated usage in the articles we analyzed into these categories: the total number of times the phrase “officer-involved” appeared anywhere in the articles; the total number of articles that used the phrase at least once; and the percentage of articles that used it at least once. In some cases, the articles directly quoted a police officer or other official, thereby amplifying the value-neutral assessment. In other cases, the phrase was used directly in the author’s own words.
These numbers were tallied on a monthly basis, which allowed us to observe the shifts in usage in the aftermath of high-profile police violence or following AP guidance and other high-profile media critiques.
Our findings illustrate how long it takes to change reporters’ habits — and their reliance on police accounts of shootings. A phrase invented by police became, over the past 50 years, commonplace in reporting and more popular over the past two decades, even amid growing skepticism of police and heightened media literacy.
As media attention toward police violence increased in the mid-2010s, so did usage of “officer-involved.” The first significant increase in monthly usage came after white Ferguson police officer Darren Wilson shot and killed Black teenager Michael Brown in August 2014, demonstrating that more coverage of police violence does not lead to clearer reporting.
In August 2014, the phrase was used 243 times, and the average usage over the year jumped to 132 times per month, nearly double any previous year.
Deray McKesson, a civil rights activist who protested in Ferguson, told us that in 2014, the media seemed earnest in its interest in covering the emerging movement against police violence, but often repeated law enforcement accounts of shootings.
“I think that a lot of the media felt obligated to have this kind of story then,” McKesson said. “And I think that especially in 2014, but even today, it’s hard for people to imagine that the government lies so easily.”
Throughout 2015, a year marked by Baltimore police killing Freddie Gray in April and the November release of the 2014 video of Chicago police fatally shooting Laquan McDonald, usage of “officer-involved” was higher than in 2014. For the year, the average usage per month was 177 times. Importantly, the AP’s Froke said that in 2015, an “editor advised AP staff in an internal note to avoid the term and instead be specific about what happened.”
In 2016, use of “officer-involved” exploded after the police killings of Alton Sterling in Baton Rouge, Louisiana, and Philando Castile in St. Paul, Minnesota, both in July. Their deaths were captured on video—Sterling was shot at nearly point-blank range outside a convenience store, Castile shot in front of his girlfriend and her daughter during a traffic stop—leading to some of the largest Black Lives Matter protests since Brown’s death in Ferguson.
Usage of “officer-involved” peaked that July, when the phrase was used 447 times and appeared at least once in 13% of all articles about police violence. That was up significantly from the previous two months; in May 2016, there were 132 uses of the phrase and in June, 100. The 447 uses of “officer-involved” in July represented the single highest usage of “officer-involved” in a month between 2000 and 2021.
Usage remained high in 2016, with the phrase appearing 228 times in August and 225 times in September. That month, the AP noted on Twitter that it had no style rule on “officer-involved shooting” but was “discussing it.” For October, November, and December, usage slightly declined, averaging 158 uses per month.
In 2017, AP Stylebook added “officer-involved shooting” to its “Cliches, Jargon” section. Along with warning reporters against using the phrase, the entry provided an example of the phrase in action: “Police say an intoxicated person of interest suffered a self-inflicted gunshot to his left foot in an officer-involved shooting.”
That year, July had the most uses of “officer-involved.” The phrase was used 152 times, appearing at least once in 14% of all articles. That month saw one of the highest-profile police shootings in recent years. A Minneapolis police officer shot and killed Justine Damond after she called 911 to report a sexual assault in an alley behind her home. The Damond case was highly unusual: She was a white victim of police violence, and the officer who killed her was the first former officer in Minnesota to be convicted of murder for an on-duty killing. (In June 2019, he was sentenced to 12 and a half years in prison on second-degree manslaughter and third-degree murder charges — in September, the Minnesota Supreme Court partially overturned his conviction and in October he was re-sentenced to 57 months in prison.)
In 2018 and 2019, usage also remained the same—on average, around 100 times per month, appearing in 13% and 15% of articles, respectively.
Then came the year with the largest civil rights uprising in a generation. In the first half of 2020, use of “officer-involved” was high — between 17% and 20% of articles used the phrase. In May 2020, the month of Floyd’s murder, usage declined to 8% of the articles surveyed and the phrase was used 130 times.
In June 2020, the month after Floyd’s murder, “officer-involved” was used 363 times.
In late August, the AP issued its guidance instructing reporters to avoid using “officer-involved” and tweeted it from its Stylebook account.
After Floyd’s murder, rates of usage of the phrase plummeted to the low single digits, but it’s unclear if that was due to simply so many more articles on police violence than was typical. However, in the months following the AP guidance, as coverage of police violence began to decline overall, the number of times the phrase was used and the rates of usage also remained low.
But progress was short-lived. In April 2021, the month former police officer Derek Chauvin was convicted of murdering Floyd, the phrase was used 136 times — but the rate at which the phrase appeared in articles remained in the 5% range. “Officer-involved” even snuck its way back into coverage about Floyd.
An April Boston Globe article about the conviction described this response from the city’s top prosecutor: “Wiping away tears, Suffolk District Attorney Rachel Rollins recounted the pain she felt watching the Chauvin trial along with other instances of officer-involved shootings that started with non-violent misdemeanors.” It was an especially strange use of “officer-involved” because Floyd was not one of several “instances” of people shot by the police — Floyd died from “cardiopulmonary arrest complicating law enforcement subdual, restraint, and neck compression” after Chauvin knelt on his neck for over 9 minutes.
As had happened after other high-profile police killings, the usage rate of the phrase declined immediately after significant events but then rose again.
Even so, rates of usage of “officer-involved” are lower now than the previous few years — they are similar to rates in the 2000s — but with many more articles about police violence, the phrase is still reaching more readers, more often.
Lowery told us he was “not surprised” by these findings. He said he was not entirely encouraged by growth in the coverage of police violence either.
“I think it’s a good thing that there was more coverage in volume but the fact that those mentions [of ‘officer-involved’] went through the roof, underscores that the median coverage is bad,” Lowery said. “If we write about it more, that is not a net good thing. Not if the writing is bad, not if the coverage obfuscates what happened and launders the police narrative into public consciousness.”
In the past year, more reporters have criticized the use of “officer-involved shooting,” especially in the wake of Chauvin’s conviction, as well as the fatal shooting of 13-year-old Adam Toledo on March 29 by Chicago police officers and the April 11 killing of Daunte Wright by a Brooklyn Center, Minneapolis, officer.
The Toledo and Wright killings were both caught on bodycam footage and released to the public, sparking nationwide protests. On April 14, Alissa Walker, a New York Magazine contributor and a co-host of “LA Podcast,” tweeted, “I was thinking last night about how newsrooms and style guides are evolving. AP Stylebook says don’t use this type of copaganda language. But lots of publications still do. Guess which publication will get access?”
McKesson said that it’s apparent that reporters and their editors — especially because much of the media is white and therefore often reporting on rather than to Black people — need to be pressured to continue to do better or they will revert to old, bad habits.
“If we went to editors and were like, ‘Can you commit to saying police shot?’ I think that they wouldn’t want to fight it. I haven’t seen editors double down, right? I’ve just seen, ‘Well, we’ve always done it this way,’” McKesson said. “They just need to be pressed.”
Lowery said that reporters responding to media criticism and grappling with how they cover police violence are crucial to changing practices.
“I think that there’s power to having the conversation in public and interrogating how we do it,” Lowery said. “The only way for that public pressure to be successful is for the conversation to be led by journalists. These critiques from within the industry carry more weight.”
Readers have a role in effecting newsroom changes too. In October, after the Los Angeles Times reported “woman hospitalized after Simi Valley officer-involved shooting,” critics swarmed on Twitter, forcing one of their reporters to admit that “It is jargon that should not be used. But we obviously haven’t instituionalized [sic] that change across a large paper. Working on it.”
In December, after a pair of Los Angeles Times headlines about Orellana-Peralta’s killing — “2 killed in shooting at North Hollywood store, including suspect,” and “14 year-old girl in dressing room killed by LAPD bullet in shooting that also left suspect dead” — garnered criticism from readers, the newspaper tweeted, “For the record: We’ve deleted a previous tweet that used passive language to describe this shooting.”
“‘Officer-involved shooting’ is an excellent meat-and-potatoes example of the way time and time and again, we sanitize violence committed by the government,” Lowery said. “Too much of journalists’ coverage of the police writ large, an arm of our government — the armed arm of our government — is too deferential and does not act as an independent entity to cover, interrogate, and contextualize an act of violence by the government.”