How to win an online debate

And in a new book, Blevins and co-author James Lee of Northwestern University say one political party used these talents and an algorithm that rewards emotional engagement to dominate the online conversation. In “Social Media and Digital Politics: Networked Reason in an Age of Digital Emotion,” Blevins and Lee use machine learning and social network analysis to examine posts on the site X, formerly Twitter, during the 2016 presidential race and the COVID-19 pandemic.

“Social network analysis shows how different groups are communicating with each other and how they are connected,” Blevins said.

Researchers examined how the tight race between Democrat Hillary Clinton and Republican Donald Trump played out on X.

“One of the first things that struck us early on is how effective Donald Trump and the MAGA movement were in this space,” Blevins said.

Blevins said social media likely played a crucial role in helping Trump win the 2016 election.

“Social media didn’t decide the 2020 election but it shaped people’s perception of it,” he said. “And I think it will shape our perception of the 2024 election, particularly if it’s close.”

Using a machine learning algorithm, researchers identified and plotted posts aligned with each candidate, audience or campaign topic to show graphically what these networks looked like and how they were connected to the larger social media space. While there were more pro-Clinton posts during the campaign, the pro-Trump network dominated the space, Blevins said.

“There were a bunch of Clinton networks, but they were isolated on their own islands. They weren’t speaking to each other,” he said.

“At the center, there was a continent made up of the pro-Trump network. They were all sharing the same themes and messages, using the same hashtags consistently. They were smaller in number but you could see the strength of their network.”

Researchers also examined the content of posts for clues to why they resonated with followers. They categorized the top users’ tweets by the rhetorical strategies they observed.

Researchers found that the political left were more likely to call on expertise while the far political right turned to personal experience and emotion.

“What Trump and Trumpism seemed to demonstrate since 2016 is the power of emotion and hyperbole in crafting political and cultural narratives against more practiced forms of reasoning,” the authors said.

Likewise, they examined posts in 2020 surrounding the COVID-19 pandemic, particularly over issues such as masking, hand-washing and controversial treatments.

“We found that when emotion was expressed in the content of Twitter posts, it tended to get shared more frequently and commented on more frequently than posts that used only science or evidence or what we might think of as reasoning,” Blevins said.

Similarly, Blevins said, public health appeals about the pandemic on social media got more engagement when they were personal rather than scientific.

Meanwhile, researchers found that traditional arbiters of fact — mainstream news outlets — barely factored into social media discussions. Individual journalists often had more influence online than the outlets they represented, Blevins said.

Lee said news organizations serve an important role in connecting disparate audiences that have their own echo chambers online.

“We found during major events, such as the COVID-19 pandemic, the news served as an information bridge connecting communities of users that normally only interact within their political tribe,” Lee said.

Featured image at top: UC Professor Jeffrey Blevins teaches in UC's Department of Journalism and the School of Public and International Affairs. Photo/Andrew Higley/UC Marketing + Brand