When the violent insurrection at the U.S. Capitol building took place, many people were surprised. But not extremism researchers such as IU’s Gunther Jikeli.
For the last several years, Jikeli, an historian and sociologist in the College of Arts and Sciences at Indiana University Bloomington, has been working with a team of graduate students to study the world of online hate speech and how it relates to actions and events in the offline world.
“You find a lot of very awful stuff, the worst things you can imagine, on the internet. We know it’s out there. The question is, what’s the impact?” Jikeli said. “How is it related to the offline world? That connection is difficult to see, but you can see it in the actions at the Capitol. We’re trying to figure out how we can better see those connections.”
Jikeli has studied online antisemitism and is the author of “European Muslim Antisemitism: Why Young Urban Males Say They Don't Like Jews.” He is also the Erna B. Rosenfeld Professor at IU’s Institute for the Study of Contemporary Antisemitism as well as associate professor in Germanic Studies and Jewish Studies in the College of Arts and Sciences.
With the support of funding from IU’s Racial Justice Research Fund, Jikeli is applying lessons learned from his online antisemitism work to analyze posts about racism online.
“Antisemitic tweets are often also racist, the content can be very similar,” Jikeli said. “In summer 2020, I realized we had an opportunity to apply methods we’ve developed already to study racism."
Jikeli and his team use Twitter data in their research, mainly provided through IU’s Observatory on Social Media. Combing through millions of tweets, the research team uses keywords and hashtags to extract a representative sample of tweets, in this case, of posts about racism.
Then they determine a detailed definition of the “-ism” being studied and go through every tweet in the sample, noting if the definition fits. The goal is to establish a “gold standard” definition that can then be applied through artificial intelligence to analyze tweets.
“Our aim is to amass a large enough annotated sample so that algorithms can be trained to detect and identify similar tweets,” Jikeli said.
Jikeli and his team get assistance by holding datathons and hackathons that engage high school and college students in working with the large dataset to annotate and identify tweets. They held an Antisemitism Datathon and Hackathon last fall and will hold a similar event for their study of online racism in summer 2021.
To examine links between online discourse and offline activities, Jikeli’s research team relies on FBI data on hate crimes, looking for statistical correlations between relevant tweets and peaks of hate incidents in particular geographic areas.
“We’re trying to examine what kinds of online speech trigger offline hate actions,” Jikeli said. “What is the role of influencers? Do certain words increase the likelihood of hate crimes? Does certain speech facilitate offline incidents?”
In the best of circumstances, this kind of social media analysis is painstaking, time-consuming work, but Jikeli said the current analysis of anti-Black racism has hit two additional roadblocks.
First is the definition of racism. Because racism is such a politicized and subjective term, Jikeli noted that it’s a challenge to determine what precise definition to apply to the data.
The second problem is the widespread social media deplatforming that followed the U.S. Capitol riots.
“Most of the extremists have left Twitter or have been banned and that presents a problem from a research perspective. We have data from earlier tweets that have been deleted, but we can’t use it. Twitter requires researchers using its data to rely only on live tweets,” Jikeli explained.
Jikeli and his team may be able to follow the extremists to encrypted channels such as Telegram, but they’ll have to change a lot of the systems they use for scraping and formatting the data.
Why keep up the pursuit into increasingly darker, more hateful corners of the online world? Perhaps counterintuitively, the reason is to find some common ground.
“People are increasingly living in separate worlds in this country,” said Jikeli, a native of Germany. “We have to find ways to come together.”
As an observer of society who’s “interested in how people think,” Jikeli said that to prevent people from adopting white nationalist thinking, we need to know and understand what arguments are drawing them in.
“If we can understand how people get more extreme, maybe we can catch them when they’re open to counternarratives,” he said.
And there’s some reason for hope, according to Jikeli. As the research team has been collecting tweets about racism, he said he’s been surprised by the number of people condemning racism and calling for a more just society.
“It’s not all bad,” he said. “You find the most horrible stuff online, but there is also a lot of good intentions and reflections.”