Really, a Twitter sockpuppet got you fired? What sounds like an absurd premise might actually not be far off the mark.
Using data science, Bethany Lacina, an associate professor of political science at the University of Rochester, found that a recent Twitter spat that ended up causing a science fiction writer to lose his job, was in part fueled by bots and sockpuppets—online identities used for purposes of deception. The automated and semi-automated accounts sowed discord, amplifying online outrage between ultra conservative and liberal Star Wars and science fiction fans on Twitter.
Chuck Wendig, by his own account, was fired as a freelance writer for Marvel’s Star Wars comic Shadow of Vader and a forthcoming Star Wars book because of the controversy his own tweets had generated.
Enter Lacina. To her, the ultra-conservative trolling that proved Wendig’s professional undoing seemed suspicious. In a study, detailed in a Vox article, Lacina dug deeper into the Twitter backlash, using Botometer, an algorithm developed by scientists at Indiana University. The Botometer allows closer examination of Twitter accounts to determine the likelihood of automated (so-called bots), anonymous, or semi-automated accounts (so-called sockpuppets).
- Read Lacina’s research as detailed on Vox.
Lacina discovered that fewer than 250 “people” (some of them also automated accounts), replied directly to Wendig’s tweets in the first 24 hours. That changed dramatically after a conservative Star Wars fan, Ethan Van Sciver, made a YouTube video attacking Wendig, whose posts are liberal-leaning.
After the video, the number of tweets from real people spiked—and so did the numbers from bot and sockpuppet accounts. About 30 minutes following Van Sciver’s tweet to his video, Wendig received about 600 tweets from legitimate accounts—and about 400 from automated and anonymous accounts, amplifying and distorting the real outrage from ultra-conservative groups, who prior to the YouTube video had typically not follow Wendig’s tweets.
“What surprises me is that these puppet and bot accounts were active in a case that didn’t get mainstream media coverage and where the person involved is not that famous,” says Lacina. “I think we’re not too far away from non-famous people using this technology against other non-famous people. It is not that much more difficult than doxing.”
Doxing, by the way, means searching for and publishing private or identifying information about a particular person on the Internet, typically with malicious intent.
Just last month, Lacina used the same big data tool, the Botometer, to analyze the tweets of angry Star Wars fans and published her findings in the Washington Post blog The Monkey Cage.
Read more about Lacina’s research on Star Wars fans’ Twitter hate posts that take aim at the franchise’s newly found diversity. She discovered that hate among fans tended to be directed more often toward women and minority characters in The Last Jedi movie, compared to non-minority characters.