Written by Jessica Wolf
In 2016 and 2017, “fake news” became one of the most influential phrases in recent memory.
But even as those two words were becoming ubiquitous, a trio of UCLA professors took a significant step toward understanding how false or misleading stories spread on the internet. Their research could lead to concrete ways to intervene — for example, by showing where and how to insert facts to combat the falsehoods.
The faculty members, representing three different schools at UCLA, set out to learn how stories get traction online. Their first target: The narrative, now disproven but still pervasive, that fed the rise of the anti-vaccination movement in recent years.
They created an algorithm to comb through millions of posts about the subject on two popular parenting websites. The program identified four key elements of the most prevalent stories — the originating event and major people involved (a family with a newborn, for example), a real or perceived threat (the vaccination), a strategy to counteract that threat (deciding to avoid vaccinating), and the success of that strategy.
In effect, we are creating a magic mirror to allow people to see the underlying narrative structures that inform any domain.
The researchers believe they can use the method to better understand how information itself is structured, which should provide clearer insights about fake news — and help organizations with relevant information for that audience create better targeted messaging that fits within the narrative.
“Our approach allows us to discover the hidden narrative frameworks that inform discussions in any arena, be it in health care, education, climate change, energy, housing or anything else,” says Timothy Tangherlini, professor of Germanic languages and literatures in the UCLA College. “It allows us to understand not only that people are talking about something, but also how they are talking about something.”
Like many of UCLA’s innovative research projects, the initiative blended a unique mix of expertise. In addition to Tangherlini, the team was made up of Vwani Roychowdhury, a professor of electrical engineering who specializes in machine learning, and Dr. Roshan Bastani, a professor of health policy and management at the Fielding School of Public Health.
The group already is planning to study how stories spread about food and phenomena such as GMOs. Their goal is to aggregate millions of internet posts to understand what spurs discussions on social media and in everyday life.
“In effect,” Roychowdhury says, “we are creating a magic mirror to allow people to see the underlying narrative structures that inform any domain.”