In an interview with the Washington Post last month, Twitter CEO Jack Dorsey said the company was experimenting with features that would promote material from opposing political viewpoints on users’ timelines, in an effort to puncture users’ self-curated information bubbles.
But research from Duke University suggests that may be a counterproductive tactic to prevent further political polarization.
In its research, published earlier this year in the journal Proceedings of the National Academy of Sciences, the team surveyed 1,200 US Twitter users who self-identified as either Republican or Democrats. They then paid them to follow a bot that retweeted elected officials and pundits from their opposing party for one month.
No political conciliation emerged. Self-identified Democrats were slightly more liberal in their views at the end of the experiment, while self-identified Republicans expressed significantly stronger conservative beliefs.
This shouldn’t come as a shock. Previous research on behavioral psychology has shown that people tend to become deeper entrenched in their political beliefs after being presented with facts that counter their positions. This may in part be rooted in confirmation bias, the quirk of human reasoning that compels us to notice information that supports our beliefs and ignore that which doesn’t. In their book The Enigma of Reason, the cognitive scientists Hugo Mercier and Dan Sperber dubbed a corollary, “myside bias,” to describe the human tendency to more readily spot flaws and holes in other people’s reasoning than we do in our own.
Their findings, the researchers point out, don’t mean that no good can come of presenting people with arguments or facts challenging their established views. It just means that it might not be very effective to do that on Twitter, the specific medium they studied—and one that only a small minority of Americans uses, no matter how vocal (or powerful) it may be.
“Readers should not interpret our findings as evidence that exposure to opposing political views will increase polarization in all settings,” the team wrote. “The findings above should not be generalized to the entire US population, because a majority of Americans do not use Twitter.”