Algorithm-curated feeds amplify like-minded information sharing more than self-curated platforms

Evidence

The basic idea is that when you are surrounded by many infected individuals, the probability that you will become infected too goes up rapidly. This is exactly what the researchers found: people with a given leaning on a contested issue were much more likely to receive information from individuals with similar views. In other words, the information diffusion process was biased towards like-minded peers. An important caveat is that this was again mostly true for Facebook and Twitter but less so for other platforms such as Reddit. What could explain this discrepancy? Well, one of the differences the authors note is that, unlike Facebook and Twitter, people can tweak the news-feed algorithm themselves on platforms like Reddit. This evidence thus suggests that filtering algorithms on social media platforms might be reinforcing echo chambers. — Sander van der Linden, Foolproof, 446

Categorization

  • QUESTION: The Reddit contrast implies users self-choice moderates echo chambers, but does user control actually reduce bias or just relocate it? Worth chasing the original study’s data.
  • XREF: connects to: Algorithmic feeds convert the open Internet into personalized echo chambers — Lee McIntyre, Post-Truth, 192; Algorithm-curated friendship narrows one’s social world into a self-reinforcing echo; Algorithmic personalization erodes the shared public space that allows surprise — Mark Fisher (edited by Darren Ambrose), K-Punk: The Collected and Unpublished…, 1287 (+2 more)

Broader Connections