Extracting the Wisdom of a Smaller Crowd fromDependent Quantile Judgments
52 Pages Posted: 5 Oct 2021
Date Written: August 18, 2021
The task of this paper is to harness the wisdom of a crowd without calibration. We propose the Gaussian process model to account for sampling and judgmental errors in quantile judgments, and assume each expert to form predictions by linearly combining various information cues, inspired by the lens model. We develop a three-step estimation algorithm to factor quantile judgments into variable profiles (independent latent cues underlying each variable of interest) and expert profiles (each expert’s weights overthese cues). We can inquire about the dependence between experts using the weightsof latent cues, because they preserve the same clustering results as the weights of actual cues up to a full-rank linear transform. After clustering experts based on their estimatedprofiles, we pick one delegate from each group. Simulation and case studies demonstratethat the selected sub-crowd can represent the entire panel well in a new elicitation task.
Keywords: Wisdom of crowds; elicitation; quantile aggregation; lens model; matrix factorization
Suggested Citation: Suggested Citation