Racial justice was a central theme of 2020, with protests and demonstrations sweeping streets across the U.S. But how did these events in response to the police killings of George Floyd and many others shape subsequent reporting of policing and crime?

That’s the central question of a new report, published last week, from researchers affiliated with the Media, Inequality and Change Center (MIC) at the University of Pennsylvania’s Annenberg School of Communication.

To compile the report, the researchers looked at a year’s worth of news coverage in three major U.S. dailies — The Louisville Courier-Journal, the Minneapolis Star-Tribune, and The Philadelphia Inquirer — in cities where there were high-profile police killings of Black citizens (Breonna Taylor, George Floyd, and Walter Wallace, respectively). The report covered stories published between January 2020 and December 2020, and compared coverage before and after Floyd’s death in May 2020. The overall sample included 830 stories across the three newspapers, nearly 75% of which were stories published after George Floyd’s death (possibly also reflecting the increased attention to police killings, police reform and protests as a result of Floyd’s death).

The authors could see the shift in trends before and after George Floyd’s death as they analyzed the material, at least with broader coverage of policing. “More in-depth stories about policing started to appear, and more contextualization also started to appear,” after Floyd’s death, according to Briar Smith, one of the authors of the report and associate director of MIC.

Among the main findings, the report found that the use of dehumanizing language — such as describing civilians targeted by the police as “suspect,” “juvenile,” or “offender” — decreased in stories published after George Floyd’s death. In the nearly five months before Floyd’s killing, such language was used in 31% of stories published in The Philadelphia Inquirer and The Louisville Courier-Journal and 27% of stories in The Minneapolis Star-Tribune. Those figures after Floyd’s death dropped to 14%, 16% and 26%, respectively.

Another example of questionable word choice in crime stories is the use of “distancing language,” including phrases like “officer-involved shooting” or “discharged weapon.” The report found that such language was “quite rare in all three newspapers, appearing in only 2% of stories in the Louisville Courier-Journal and 6% of stories in the Minneapolis Star-Tribune and Philadelphia Inquirer.” (The authors did not provide a comparison between the usage of distancing language before and after Floyd’s death.)

As far as what could explain these findings, “There was already some awareness about the need to change some basic practices, like mug shots, for example,” said Susanna Dilliplane, who is deputy director of The Aspen Institute’s Planning and Evaluation Program, but who worked on this report as an independent consultant. “Dehumanizing and distancing language have gotten a little bit more attention because they were kind of a lower-hanging fruit.”

The report also found that in the period following Floyd’s death, mention of racial justice for contextualizing stories about policing increased significantly across all three newspapers. Nearly 30% of crime-related stories — and 34% of protest stories — in the Star-Tribune had references to racial justice from May 25, 2020 onwards. Previously, only 6% of crime stories specifically referred to the racial justice context.

Those figures were higher for the other two papers. More than three-quarters of protest stories in the Courier-Journal and the Inquirer included the racial justice context, and nearly half of all crime stories in the Courier-Journal included this context following Floyd’s death (up from 10% in the first five months of the year).

That’s the good news. There were other content-specific findings in the report to suggest that local crime stories (versus stories about police conduct, accountability, and reform in general) were less likely to publish stories that challenged the police.

For instance, only 5% of local crime stories in the Star-Tribune challenged the status quo, compared to 78% of general stories about police reform and conduct. And even though nearly a quarter of local crime stories in the Courier-Journal challenged police practices, an overwhelming 92% of other stories on reform and protests challenged law enforcement’s authority.

While a minority of stories across the three papers used police-only sourcing for their stories or the mention of police in the headlines, whether local crime or police reform in general, local crime stories tended to indulge in this practice more frequently. For example, while 23% of all policing-related coverage at the Inquirer relied solely on police as sources, but that number for local crime stories was 38%. Only 11% of all policing coverage at the Courier-Journal cited police in the headline, but that jumped to 31% among the paper’s local crime stories.

The bias towards the status quo was also apparent in the way newspapers tended to cover protests, the report found. Across all three newspapers, an overwhelming majority of stories about protests portrayed protesters as posing a threat to property, to police officers, and to other civilians. And the proportion of stories with this kind of portrayal far exceeded the proportion of stories that mentioned harm to protesters because of police (due to the use of tear gas and rubber bullets, for instance). The proportion of stories in the Star-Tribune outlining threat posed by protesters was more than three times the proportion of stories in that paper outlining protesters harmed by police, for instance.

The report also found paper-specific trends, such as how The Philadelphia Inquirer was the most likely to publish stories with only police as sources and how the presence of police sources in the Minneapolis Star-Tribune’s stories actually increased following George Floyd’s killing.

Beyond takeaways for the individual newspapers included in the study, the report also includes advice for other newspapers looking to improve their coverage of police- and crime-related stories, and in some cases, for other beats as well.

“Sourcing to me is a no-brainer,” said Dilliplane about the opportunity for newspapers to diversify who is quoted in stories, beyond authority figures. “I think there’s just broader recognition that more inclusive sourcing practices are needed in order to really accurately reflect and represent the experiences of communities but especially ones that are misrepresented or maligned by the mainstream, white-led media,” she said.

Other implications range from reevaluating how responsibility is framed within a given story (such as mention of a civilian’s criminal history or shifting responsibility away from police) and addressing the imbalance between coverage that emphasizes the threats (real or perceived) posed by protests versus police actions.

It also helped to be able to see the breakdown between coverage practices in general policing stories versus local crime stories, especially since broader stories about policing seemed to be moving away from traditional tropes of crime reporting. “We see some evidence of more inclusive sourcing practices in these other kinds of stories, [and so] maybe there’s a way just to expand those positive examples into some of the local crime coverage,” said Dilliplane. “You can build on what already seems to be moving in the right direction as perhaps a response to last year’s [protests].”

There may be lessons for everyday reporting as well. The report underscores the importance of giving reporters time to properly report, Dilliplane said. “I think one of the reasons you get some of this more superficial stereotype-reinforcing coverage is because they don’t have time to to do the kind of follow-up that they need to do or to chase down the sources or build the community connections they need to in order to reflect those more diverse voices.”

This issue is likely to get more attention. Just last week another study from the Garrison Project found that use of dehumanizing and distancing language has steadily increased in the past two decades, and that guidance on moving away from such language hasn’t really made an impact. This study’s findings further reinforce the call for journalists and newsrooms “to critically examine patterns of coverage that fundamentally shape how readers think about and understand policing in their communities,” Louisa Lincoln, a PhD student at the Annenberg School for Communication and another author of the MIC report, said. “As both of our studies clearly illustrate, newsroom practices have real and lasting consequences on the narratives that emerge from incidents of police brutality.”