Nov 9, 2017 - Economy

How news can drive our conversations

Illustration: Rebecca Zisser / Axios

News stories have a measurable impact on Americans taking to Twitter to talk about policy issues, according to a study published today in Science. Researchers found people — regardless of their gender or political affiliation — discussed race, immigration and other topics more often after stories were published than if news outlets weren't covering the issues.

The bottom line: If a few small outlets can have an effect on the national conversation as the study suggests, study author Gary King says bad actors may also be able to have a big impact. "And so we all have this responsibility to make some kinds of decisions about the entire ecosystem since it seems to be highly influential," according to King, who's a social scientist at Harvard University.

Be smart: This isn't about the power of fake news but of how media can be used for propaganda. "Most people are missing the fact that propagandists have long followed the rule that they should always tell the truth. The reason is that if they say something false and get found out they lose credibility. A much more effective propaganda strategy is to tell the truth but to amplify the arguments you like," King tells Axios.

What they did: The research team recruited 33 media outlets and advocacy blogs — largely digital and mostly small, ranging from Defending Dissent to The Nation to the Huffington Post. Between October 2014 and March 2016, "packs" of 2–5 outlets wrote and published articles on 11 different policy areas, including race, immigration, climate and education. They asked the groups to choose a specific subject within the area for all members to write about and they then ran the stories at the same time.

The researchers then tracked Twitter discussions on the same topics and reported that, on average, the stories published by a group increased social media posts about a policy area by 10% — or 13,166 additional posts — over the week following publication compared with what was normally being tweeted on the topic. The researchers said most people didn't refer to the story per se but to the larger issue.

One interesting thing: The effect was the same regardless of political party affiliation, geography, gender or how many Twitter followers people had. "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," the authors wrote.

Keep in mind:

  • The interventions took place during quieter news periods. Major news events could change the effect, the authors noted.
  • Policy pales when compared with entertainment: For example, the team's policy topics generated 1/100th of the tweets circulating when a new episode of Scandal was about to air.
  • Most importantly, the study looked at relatively small news organizations compared to the New York Times or Fox News. The researchers did try to get a handle on the impact of an outlet of that size. They looked at the effect of a NYT article about fracking's impact on drinking water — published when the topic wasn't being widely discussed — and found a 300% increase in people talking about water quality in the day that followed (versus 19% in their experiment).
  • It's also tough to say on an individual level whether someone's focus on a subject is coming from the news or their own interests, says University of Maryland economist Ethan Kaplan, who was not part of this study.

The big question remains: Just how much of an impact does news media have on what voters or legislators do? One interpretation of the study is that it is a short-lived effect and so politicians don't have to care about reporting on an issue, says Kaplan, who has studied Fox News' influence on voters. "Or, it could be that it is changing someone's view on these issues — they just don't tweet about it as much or they go ahead and vote differently."

Go deeper