A 2018 change in Facebook’s algorithm resulted in significant boosts in engagement for local Republican groups even though their Democratic counterparts posted more often, according to newly published research. The findings, first reported byNBC News, line up with one of the major revelations of the Facebook Papers. Namely, that a change meant to emphasize content from family and friends wound up making News Feed more divisive by incentivizing negative posts.
The latest research, published in Research & Politics, used CrowdTangle data to measure engagement with posts from local Democratic and Republican groups on Facebook and Twitter between January 2016 and August 2021. The researchers found that by the fall of 2018, several months after Facebook announced its algorithm change, there was a significant uptick in engagement with Republican pages on Facebook that didn’t occur on Twitter.
Of course without the actual algorithm we cannot say what caused the change in 2018. We can say there was a change, specifically on Facebook, and that it impacted a large number of local Republican parties. We can also say the timing lines up with what others saw for FB's changes
— Kevin Reuning (@KevinReuning) June 8, 2022
“We conclude that it seems possible that changes in how Facebook rated content led to a doubling of the total shares of local Republican party posts compared to local Democratic party posts in the first half of 2019 even though Democratic parties posted more often during this period,” the paper’s authors write.
The researchers say it’s “difficult” to know for sure what caused the shift toward the end of 2018. But they note that their findings broadly align with what we know about Facebook’s algorithm change, and the reaction to it, thanks to the Facebook Papers. The company had announced a major change to News Feed’s ranking systems in January 2018, in order to emphasize “meaningful social interactions” or MSI. But documents disclosed by Facebook whistleblower Frances Haugen indicate that by early 2019 publishers and political parties were raising questions about whether MSI was incentivizing negativity.
“Of course without the actual algorithm we cannot say what caused the change in 2018,” Kevin Reuning, one of the paper’s authors wrote on Twitter. “We can say there was a change, specifically on Facebook, and that it impacted a large number of local Republican parties. We can also say the timing lines up with what others saw for FB's changes”
In a statement to NBC News, Meta called the researchers’ findings “implausible,” saying differences in engagement could be attributed to other factors. “It doesn’t add up with what MSI actually did, which was reduce the amount of public content — like that of political parties — on the platform,” a company spokesperson said. “The trends here instead seem to coincide with a divisive election cycle, and since the differences between political parties in the U.S. have been growing for decades, the idea that a change to Facebook ranking would fundamentally shift how people choose to engage with political parties is implausible.”