Daniele Vilone and Eugenia Polizzi (2024), PLOS ONE 19(1): e0296075.
Key takeaways
From the reading, NOTE-810.
- Beliefs fixed, expression free. Agents hold a private belief ±1 that never changes, and choose only what to express. Any misperception is therefore about the distribution of opinion, not a change in it.
- Topology alone does little. The distortion of the local climate relative to the global expressed opinion is “at most around 15%”. In the second model it even moves perception toward the true beliefs. The authors call it “besides small, also non-universal”.
- Silence does most of the work. When agents may stay silent, each picks the expressed value closest to a degree-weighted mean of its own belief and its neighbours’ expressions. About 55–60% of agents, almost all of the majority, end silent, and they do so “regardless of K_c”.
- Number and centrality trade off. A large, poorly connected minority and a tiny, highly connected one silence about the same share. At K_c = 20 each hard-core agent silences about 150 others, well beyond its neighbours.
- No reversal below about 30%. The minority looks bigger than it is, but overturns the expressed majority only when it exceeds about 30% of the population.
- Simulation only. The authors call for validation against data on private and expressed opinion.
Standing in the record
Filed on 2026-10-12 for QUESTION-047 and read on 2026-10-14 in NOTE-810. What it shows is held as THEORY-242. It bears on the question’s “which community-level mechanisms”, but for what a community appears to hold, not for a belief it settles on. The seeded claim that “the outcome is set by structure, not by any member’s reasoning” was half wrong. The paper finds the direct structural effect small, and locates the bias in the agents’ expression rule. Structure enters through degree-weighted influence and the cascade.