When small news outlets simultaneously publish articles on broad topics such as jobs, the environment or immigration, the stories can prompt people to talk more about such major issues of public policy and politics, according to a study published in the November 10 issue of Science.
A team of social and political scientists measured this increase in public discourse by tallying the number of social media posts on these topics in the week after the articles’ publication. The increase in social media posts held true across different political affiliations, geographies and other subgroups, highlighting how the role of U.S. journalism remains highly influential across a broad section of people — more so than many have thought.
“Apparently, the national conversation really is one conversation, at least among those able to participate in social media; even if they do not interact with each other, the evidence indicates that they are being influenced in similar ways by the news media,” Harvard University professor Gary King and his colleagues write in the Science paper.
The findings demonstrate that the news media can wield considerable power in generating national dialogues that are essential to a healthy representative democracy, the researchers conclude.
“The media establishment is changing fast,” said King, Albert J. Weatherhead III University Professor at Harvard, citing declines in the number of journalists, as well as changes in the business models of media organizations — including, for instance, shifts from revenue generated through print advertisements to revenue through subscription paywalls.
“Amidst these changes, journalists are doing their jobs, trying to have an impact,” King said. “We showed that they indeed do have a big impact on the content of the national conversation; even journalists at smaller outlets can affect what Americans talk about nationwide.”
Estimating the influence of the news media on public expression — on whether it causes individuals to take public stands on key policy issues, for example — has long been challenging to scientists.
“For several hundred years, researchers have been trying to measure the influence of the media,” said King. “We always thought it was influential, but this is a question that’s very difficult to study.”
King’s team has been working on ways to study this problem for more than five years, much of which was spent convincing various news outlets to agree to take part in a randomized study.
“Our goal was to not interrupt journalists’ daily practices,” King said. “We tried to fit into their standard operating procedures, and not to ask for compromises from their standards.”
The researchers worked with news outlets to develop a model similar to the oversight and data sharing practices used by multiple news organizations in the collaborative publication of stories based on shared documents such as the Panama Papers in 2015 or the recently released Paradise Papers. The news outlets were free to choose to participate in each experimental cluster of stories. When the outlets joined an experimental cluster, they developed and published specific stories under the broader topic area chosen by the researchers.
“Apparently, the national conversation really is one conversation, at least among those able to participate in social media; even if they do not interact with each other, the evidence indicates that they are being influenced in similar ways by the news media,” Gary King
“[Our design] enabled both full randomized experimental control in the hands of the researchers,” King said, “and full editorial control in the hands of the journalists, fitting into their familiar customs and practices.”
“No one has been able to randomize this type of study like we did,” King said. “That made it possible for the first time to run a large-scale media experiment.”
King and his team recruited 48 mostly small news media outlets, chosen based on circulation size and Google Analytics pageviews. These included Cascadia Times, Defending Dissent, News Taco and Big Education Ape, among others. Researchers also recruited a few mid-sized mainstream outlets, including The Nation and Public Radio International, and one large national outlet, Huffington Post. Next, King and colleagues chose groups of two to five of these outlets to write and publish articles on broad policy areas such as immigration, climate and education.
Within these broad topic areas, news outlets developed stories of multiple types from investigative stories to features and stories based on interviews. “For example, if the area was technology policy, the story might be about Uber drivers in the Philadelphia area,” King said. “We would then try to identify a two-week period where we predict there wouldn’t be any surprises related to that topic area — so if the president was going to give a speech about immigration, we would not run an experiment [on that topic] during that time.”ad
Each cluster of stories published by the outlets on the same topic was designated to run on one of two consecutive weeks.
The study authors then assessed the impact the story cluster had on prompting social media conversations by comparing outcomes in the “treatment” week, in which the cluster of stories ran, to the “control” week, in which the stories did not appear. They used tools developed by a Harvard-based start-up Crimson Hexagon to monitor social media posts.
In the week following publication, the average cluster of stories generated more than 13,000 additional Twitter posts by more than 7,000 unique authors on the topic, compared to social media posts on the topic during the control week — an effect larger than anyone anticipated, according to King.
The power of the media to generate discussion regardless of readers’ political affiliation, gender or geographical location suggests that the polarizing effect of new media outlets such as Twitter could be more limited than some have thought, said Matthew Gentzkow, a professor of economics at Stanford University, in a related Policy Forum .
Notably, the percentage of social media opinions that agreed with viewpoints expressed in the story clusters increased by 2.3% after the stories were published, the researchers say. For instance, articles on demonstrations supporting the DREAM Act, a proposal to grant residency for minors illegally in the United States, could increase the proportion of pro-immigration posts after publication of the articles.
King and colleagues say more work is needed to understand how the media influence individual attitudes and persuasion and call for further studies to determine the impacts of media ownership on influencing national policy conversations.
“Given the tremendous power of media outlets to set the agenda for public discussion, the ideological and policy perspectives of those who own media outlets have considerable importance for the nature of American democracy and public policy,” the researchers write in their Science report.
“The conversations [we evaluated] are essential to American democracy,” King concluded. “They suggest that every journalist has a huge responsibility.”