Facebook’s own research measured ideological alignment by averaging sharers’ self-reported politics

Evidence

Although Facebook admits that ‘provocative content’ leads to more engagement on the platform, like YouTube, they are sceptical about how much this actually contributes to the formation of echo chambers and filter bubbles. In fact, a couple of years ago, Facebook decided to weigh in with their own study, published in the prestigious journal Science. Eytan Bakshy (a Facebook data scientist) and his colleagues looked at a set of 10 million active US Facebook users who self-reported their political ideologies and shared about 7 million distinct URLs (web links) over a six-month period. They determined whether a news story was congenial to a particular ideology by averaging the ideological affiliation of all people who shared a given story. So, for example, a story from Fox News would most likely be shared by conservatives and would thus receive a high alignment score with conservatism as opposed to liberalism, and vice versa for a story from CNN. — Sander van der Linden, Foolproof, 417

Categorization

• QUESTION: The method of inferring a story’s ideology from average sharer affiliation is coarse — a possible critique worth chasing for how platform studies measure echo chambers.
• XREF: connects to: Survey shifts in ideological self-identification measure symbolic meaning, not issue movement. — John T. Jost and Jim Sidanius, Political Psychology, 2081

Broader Connections

• Survey shifts in ideological self-identification measure symbolic meaning, not issue movement. #qualifies (0.85) — John T. Jost and Jim Sidanius, Political Psychology, 2081