MIT Sloan study identifies political disagreement as a strong motivator for fact-checking tweets
CAMBRIDGE, Mass., May 18, 2022 – Stopping the spread of misinformation while maintaining free speech is a major challenge for social media companies. Elon Musk purchased Twitter to “adhere to free speech principles,” however it is unclear what that will mean for the platform’s future attempts to fight misinformation. One approach, which Twitter was working on prior to the sale, may appeal to Musk: Rather than removing misleading content from the platform, the Birdwatch program allows participating users to provide additional context for potentially misleading tweets. A recent study by MIT Sloan School of Management and researchers Jennifer Allen and Cameron Martel found that while participants in Birdwatch are politically motivated, the program nonetheless has potential to help address the misinformation problem.
“We need to understand the role of partisanship in social media interactions to improve online platforms. If unchecked, partisanship can lead to behavior like toxic political discourse and harassment of anyone with opposing views,” says Allen.
In their study, Allen, Martel, and Rand leveraged data provided by Twitter about Birdwatch. This crowdsourced fact-checking tool allows a select group of pilot participants (“Birdwatchers”) to identify tweets as misleading or not, write free-response fact-checks of tweets, and evaluate the quality of other people’s fact-checks. Birdwatch is not focused exclusively on political misinformation, and users may fact-check any tweet they choose.
They found a strong pattern of Birdwatchers seeking out and disproportionately flagging tweets written by people from across the political aisle. Content from counter-partisans was judged as more misleading and unhelpful far more often than content from co-partisans. For example, Democrat Birdwatchers were 3 times more likely to fact-check tweets made by Republican users compared to Democratic users, while Republican Birdwatchers were 1.5 times more likely to fact-check tweets made by Democratic users compared to Republican users.
Similarly, Democrats rated other Democrats’ fact-checks as helpful 83.1% of the time, compared to 43.3% of Republicans’ fact-checks. Republican Birdwatchers rated fact-checks from fellow Republicans as helpful 87.1% of the time compared to 25.9% for Democrat-written fact-checks.
“The preferential flagging of counter-partisan tweets doesn’t necessarily mean that Birdwatch is doomed. If people mostly flag false counter-partisan tweets, then the system could actually work well. The two sides of the aisle would effectively be policing each other’s content,” says Rand, noting that a preliminary investigation of the quality of the ratings is consistent with this possibility. The authors had professional fact-checkers evaluate 57 of the tweets flagged as misleading by Birdwatchers and found that 86% of the tweets were also rated as misleading by at least one of the professional fact-checkers.
Martel adds, “Political partisanship is a good incentive for people to fact-check content. This motivation should be considered when designing platforms to encourage more accurate posts as well as more fact-checking of content. It will be interesting to see what Twitter and other social platforms do next.”
Martel, Allen, and Rand are coauthors of “Birds of a feather don’t fact-check each other: Partisanship and the evaluation of news in Twitter’s Birdwatch crowdsourced fact-checking program.”
About the MIT Sloan School of Management
The MIT Sloan School of Management is where smart, independent leaders come together to solve problems, create new organizations, and improve the world. Learn more at mitsloan.mit.edu.