Prelec, D. (2004). A Bayesian truth serum for subjective data. Science 306(5695): 462–466. doi:10.1126/science.1102081.
Not read. The summary is drawn from the abstract. A reading should check it first, and in particular the conditions (a large population, a common prior, respondents who treat their own answer as evidence about others’) under which the abstract’s claim holds.
Why this is here
Filed on 2026-10-24 with Miller, Resnick and Zeckhauser’s peer-prediction method (LIT-1518) as the other answer to scoring evaluators without an objective metric. Its criterion, answers more common than predicted, is aimed at the consensus bias that makes agreement-based scoring entrench the majority, which is the capture risk the lead named against CASE-155’s self-scoring term. Unread, it supports nothing here yet.