Our students tried a bunch of possible solutions for news bubbles. The results weren’t great.

Trying to get people out of their news bubbles? Here for it.

Actually succeeding? Not so much.

As news audiences continue to fragment into like-minded splinter groups — a phenomenon which facilitates the spread of distorted or fabricated information — more organizations are trying to fight back against the “echo effect.” USC journalism students tested several of these solutions to see how they worked from the user perspective. (You can see videos of their demos on the Annenberg Media Solutions playlist on YouTube.)

The results were less than encouraging. Although they appreciated the sentiment, the students found most of the attempts moderately helpful at best. Some had technical glitches, or weren’t robust enough to be useful. Others required too much effort from the consumer, or just didn’t do what they set out to accomplish.

Take Fiskkit, a platform which allows readers to annotate news articles with their opinions and call out facts they see as questionable. It’s a cool concept, but as Carlin Pappas found in her unscientific survey of her fellow college students, most of them aren’t that motivated to work so hard for their news. (In a comment published on the Annenberg Media YouTube channel, Fiskkit said the product is still in beta, that they’re open to feedback and that they can develop “statistically significant” analytics with a relatively small number of users.)

Daniela Pieche tried BuzzFeed’s experimental “Outside Your Bubble” modules, which display comments from Twitter, Reddit and other platforms on selected articles. They’re designed to give readers a window into social media conversations that would otherwise be happening in separate silos. She liked the feature — once she found an example. But she wanted those different perspectives on a wider variety of stories, and she would have preferred more facts and fewer opinions. It’s also worth noting that Pieche hadn’t heard about “Outside Your Bubble” before we started researching solutions — despite being an active news consumer and a journalism student.

Then there’s Twitter, which has a serious bot problem. One study found as many as 19 percent of users tweeting about politics in the month leading up to the elections were actually bots. That matters because phony users with an agenda can create the illusion of widespread support for false or misleading stories.

Madeline Ottilie tested a tool called Bot or Not, which analyzes Twitter users to determine how likely they are to be bots. It worked pretty well on institutional accounts, but had some misfires, like the account of a classmate which showed up as 61% bot. (The organization behind the effort has since released an updated version called Botometer.)

Charlotte Pruett checked out The Odyssey Online. The site’s goals are lofty — editors say they want to give more people the chance to publish work important to their communities. But some of her peers, including one former writer for the site, thought it focused on stories that were shareable rather than memorable. (A company representative said it tries to encourage its writers to reach the widest audience possible.)

Polar News also has a worthy aim. It curates opposing perspectives on news stories, and delivers them to readers via email and on its site. Judith Nwandu thought the service had some good points, but also felt she’d be unlikely to use it more than occasionally.

Some of these ideas may still take off, especially after their next round of fixes. But they also face real challenges. Anti-bubble tools can’t succeed if they only appeal to people who are already highly motivated, engaged, and willing to put in extra effort. Just like the rest of the journalism industry, they need to give audiences something that’s relevant, meaningful and easy to incorporate into what they’re already doing. Until then, the solutions movement will be more of a well-meaning hope than an effective weapon in the fight against disinformation.

Polar News

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.