Bennett Holman & Justin P. Bruner (2015), Philosophy of Science 82(5):956–968 — DOI-10.1086/683344
Key takeaways
Registered, not read; only the abstract was seen, on the authors’ institutional portals. By the abstract, it extends Zollman’s network-epistemology model with one epistemically impure agent among honest truth-seekers, and concludes that an intransigently biased agent keeps the community from ever converging on the truth. It explores two solutions, one a procedure for endogenous network formation in which agents choose whom to trust, and applies the result to medical research. The abstract does not say what the other solution is.
The authors’ 2017 paper (LIT-744) describes it as having used Zollman’s model “to investigate the medical research community”, and cites it among works suggesting that industry-funded research be given less weight. The 2017 model differs: no agent there is intransigent or swayed by its funding, and the community’s bias comes from which methods are funded.
Standing in the record
Filed on 2026-10-05 at the owner’s request. The record’s note on Rosenstock et al. (NOTE-559) reports that paper citing it for the finding that connected networks resist biased agents better; that is second-hand and unchecked here. The account read in its place is THEORY-146.
Note of 2026-10-21: a later model registered
A25 of the exchange of 2026-10-20 ranked Weatherall, O’Connor and Bruner’s “How to Beat Science and Influence People” (BJPS, doi:10.1093/bjps/axy062) among its first four works. It builds on the Holman and Bruner model, and is registered as LIT-1341, with this entry as its precursor.
It was read in full the same day (NOTE-1038). Its footnotes 14, 18 and 23 describe this work’s model: the biased agent draws from a biased arm and shares with the scientists, not with an outside audience; with such an agent among them it is best for the scientists to be highly connected; and scientists who can ignore statistically unlikely data do better when n is large, because the biased agent is easier to detect. These are Weatherall et al.’s reports, second-hand, not a reading. The standing is unchanged: Unavailable until a copy is supplied.