UW researchers race to trace online rumoring

Kate Starbird and her team's research on the response to the Boston Marathon bombing on Twitter revealed details about how rumors spread online. 

Surfing the web is a dangerous sport. Waves of distraction, riptides of doubt, and storms of debate threaten to overwhelm the average Internet user.

When real-life disasters occur, this experience is only intensified. Nowadays, more people, including emergency responders and affected individuals, are turning to social media to find and share information.

For those users purposefully seeking accurate and reliable information, it’s sink or swim.

Kate Starbird, UW assistant professor of Human Centered Design and Engineering and director of the Emerging Capacities of Mass Participation Laboratory, is trying to provide a life jacket.

Since June 2013, Starbird and her team of students have waded through more than 10 million tweets from the Boston Marathon bombings in order to better understand the spread of misinformation on social media platforms. Their research, which will be published in March 2015, details the uncharted metrics they developed for identifying and tracing online rumoring behavior. Eventually, the team hopes to develop tools that can automatically detect, and maybe even deter, these rumors.

The project began April 15, 2013, when three people were killed and an estimated 264 others were injured after two bombs exploded during the Boston Marathon. Chaos ensued, both in-person and online, as the world struggled to grapple with this tragedy.

“After a disaster event, all the information isn’t known yet, so people are trying to make sense of what happened,” Starbird said. “It leads to us coming up with explanations and sometimes those explanations turn out to be right. Sometimes they turn out to be wrong. Most of our rumors, we can see initially people are trying to make an explanation but then some bad apple comes in, purposefully.”

Such was the case for the rumor involving the death of a little girl running in the marathon. Although it was correctly reported that an 8-year-old was killed on the sidelines, it wasn’t until a Twitter user attached this news to an uncredited photo of a little girl running the marathon that the story, now altered, began going viral, according to Starbird.

The researchers also focused on other rumors that incorrectly claimed the death of a woman right before her boyfriend’s proposal, labeled Brown University student Sunil Tripathi as a potential suspect, and described the orchestrated involvement of the Navy Seals.

Unsurprised by the presence of online rumors, Starbird and her fellow researchers were curious to learn about the presence of a so-called “self-correcting crowd.” Although they had heard that this “crowd” would vigilantly attack most rumors, the data showed the opposite to be true. In reality, the marginal amount of correcting tweets were vastly drowned out by a sea of misinformed tweets, many of which continued to spread, even after being corrected.

Undergraduate Jim Maddock, who has been on Starbird’s team the longest, shares the team’s long-term goal of building a Twitter tool to help users think critically before posting misinformation. The tool would work by computationally classifying something that is spreading virally, share with the user that the information they are about to share is unverified, and prompt them to research further before spreading what may or may not be a rumor.

Starbird chose to use Twitter over other social media sites for its vast, free, and diverse data set.

Ken Rufo, a professor in UW’s Communication Leadership program adds that the site is an especially interesting outlet to explore how rumors are spread.

Rufo, who was not involved in Starbird’s study on the Boston Marathon bombing, suggested that Twitter’s asymmetrical relationships and potential for anonymity might encourage more rumor spreading than other social media sites.

“It’s slightly more pure in the sense that they fear less judgment or context that might reflect on their identity,” Rufo said.

It is hard to pinpoint the exact variables that promote online rumoring. Yet it’s a challenge that Starbird and her team are excited to tackle.

“While this research on rumoring often focuses on negative aspects of online behavior, most of the online activity I see after disaster events is very ‘pro-social,’ with people coming together to try to help each other in a multitude of different ways,” Starbird said. “It’s very uplifting, actually.”

Right now, the team is busy exploring Twitter conversations surrounding the Ebola crisis.

The researchers will continue studying rumoring behavior on social media platforms following disaster events through a two-year grant from the National Science Foundation.

Reach contributing writer Christy Pham at development@dailyuw.com. Twitter: @pham_christy

(0) comments

Welcome to the discussion.

Keep it Clean. Please avoid obscene, vulgar, lewd, racist or sexually-oriented language.
Don't Threaten. Threats of harming another person will not be tolerated.
Be Truthful. Don't knowingly lie about anyone or anything.
Be Nice. No racism, sexism or any sort of -ism that is degrading to another person.
Be Proactive. Use the 'Report' link on each comment to let us know of abusive posts.
Share with Us. We'd love to hear eyewitness accounts, the history behind an article.