Talia Stroud of the University of Texas says she first became interested in how people choose their news sources in 2004. She was beginning to sense the rise of media polarization, so she started doing some research on why people seek out news from sources that are likely to confirm their own views — and what the implications are for American democracy.

In 2011, Stroud published Niche News: The Politics of News Choice, which confirms the idea of “partisan selective exposure” and seeks to use this tendency toward self-selecting polarization to explain the interrelated workings of the media and the American political system.

“The finding that makes me the most afraid,” says Stroud, “is the people who are most likely to polarize and look at like-minded media and exhibit some of these behaviors that I don’t think are pro-democratic are those who are most politically knowledgeable.”

So what to do? One strategy, taken up by some journalists, is shaming: complaining loudly about how awful it is that conservatives only watch Fox News and read Drudge or that liberals only watch MSNBC and read Daily Kos. But while that may feel nice to the complainer, it doesn’t do much good to change behavior. Stroud wanted to try for something more actionable. The result is the Engaging News Project at UT, which received funding from The Democracy Fund, a project of the Omidyar Network via the New America Foundation, and which Stroud presented at the Personal Democracy Forum in New York last week.

From the outset, Stroud resisted the temptation to come at the problem with a mind to create a newsroom revolution. Instead, she was looking for practical, easily implemented, small-scale solutions that local media partners could begin using and testing right away.

“I wanted to develop something that would have two goals — one of which is a democratic goal and a second which is a business goal,” she said. “That’s quite doable in today’s media environment.”

Stroud and her team started brainstorming tools that would not only encourage readers to engage in more diverse discourse online and expose themselves to new viewpoints, but which also had a chance of enhancing revenue for news outlets who already spend lots of time worrying about how to get more eyes on their pages for longer periods of time.

“Before I started this project, I talked with a lot of people in a lot of newsrooms, from editors down to those who were like, ‘Hey, I’m a social media intern.’ There are some people who are only interested in the business angle, and there are some that are truly passionate about a democratic angle,” says Stroud. “I think the Engaging News Project is well positioned, because it can actually speak to both of those.”

The team located four different features of online news sites that they wanted to experiment with — comments, polls, link framing, and buttons. The results of the comments experiment — which looks at the role journalists play in guiding conversation through questions and interaction — aren’t in yet, but the other three yielded provocative findings.

A lot of news sites out there are using polls on their homepages to increase user engagement. Stroud asked whether there was a way to modify this trick to both make the polls both more informative to the reader and more interesting, so that they’d stay longer. Instead of asking questions about policy preferences — the results of which are likely unrepresentative of the broader public, anyway — she found readers learned more from being asked questions about existing poll data. In other words, asking “What percentage of the public approves of gay marriage, according to a recent Gallup poll?” generated more learning than just asking “Do you approve of gay marriage?”

The Engaging News Project also discovered that readers were more likely to engage with slider polls than other designs, and that they were actually willing to answer a series of poll questions, which is good business news for local media.

The Engaging News Project’s other two areas of focus were based on the differences a subtle change in language could make when trying to get readers to think about consuming a different kind of news.

One idea was to eschew the “Like” button — which Stroud says news orgs adopted from Facebook with little contemplation — and try to come up with a word that allows for greater breadth of meaning. After a bomb scare at UTAustin a few years ago, one of Stroud’s students saw a local TV channel encouraging viewers to ‘Like’ their coverage online, saying, “We know this doesn’t mean you like the bomb scare.”

Stroud, acknowledging at the limitations of the Like button, decided to test a “Respect” button:

From a business angle, respondents seeing a “Respect” button clicked on more comments in a comment section. From a democratic angle, respondents seeing a “Respect” button clicked on more comments from another political perspective in comparison to the “Recommend” or “Like” buttons.

Unfortunately, the feature Stroud was most excited about experimenting with turned out to be the greatest disappointment. After combing through the political science, communications studies, and psychology literature on motivating language, she came up with phrases meant to encourage people to click news links over entertainment links, and also to click links with which they might not agree.

They tested a number of phrases. The most successful: “Form accurate positions by reading different viewpoints” and “Thanks for keeping up with the news. Be proud of protecting your democracy.” But none of the lines produced replicable or reliable results.

“Some of these phrases led site visitors to evaluate a site more positively, but also decreased the number of clicks on a site,” Stroud wrote. “Others encouraged some visitors to spend less time with counter-attitudinal editorial content.” Ultimately, she said the takeaway is that the effects of language can be highly variable, so it’s best to plan carefully and test thoroughly. Stroud says she’s planning more link-framing experiments in the future in hopes of nailing down more tangible results.

“My hope is that novel tools and novel buttons can shake people out of their habits,” says Stroud. Soon, the Engaging News Project hopes to make all of these tests available as a front-end plugin for news sites, so that lots of outlets can contribute data from A/B testing without hands-on engagement with university researchers. Stroud says a WordPress plugin for a Respect button is already available.

While Stroud is hopeful that ongoing research will yield further approaches to news design that serves democracy and business equally, she’s also begun to notice an inverse relationship between passion for a cause and thoughtful conversation. For newspaper owners, the tension is between a dislike of comment-section screeds and the knowledge that publishing them will get the outlets the clicks they need. But Stroud sense another problem as well.

“Diana Mutz says there’s either participatory democracy or deliberative democracy, and you kind of have to choose,” she says. “Some people have critiqued that and said maybe we don’t have to choose. But I think she’s on to something. It’s energizing to get people to feel passionately about a cause, to go out and do anything for it. But to get them to sit back and think about all the arguments of the other side takes some of the passion out of it. It’s a tricky thing to break through.”

The Engaging News Project’s findings will be published as white papers and presented at conferences throughout the summer.

Image by Sean MacEntree used under a Creative Commons license.