Intuitions about the epistemic virtues of majority voting

Abstract

The Condorcet Jury Theorem, along with empirical results, establishes the accuracy of majority voting in a broad range of conditions. Here we investigate whether naïve participants (in the U.S. and Japan) are aware of this accuracy. In four experiments, participants were provided with information about an assembly voting to decide on one of two options, one being better than the other. In Experiments 1 and 2, participants were provided with specific parameters and they vastly underestimated the probability that the majority would select the right option. In Experiment 3, participants were provided with less specific information, and still underestimated the probability that the majority would select the right option. In Experiment 4, participants were asked to compare majority rules and supermajority rules. Most participants failed to grasp the relative weakness of supermajority rules. Our results are compatible with participant relying on a simple model of the voting situation based either on the competence of an individual voter, or on the minimum proportion required for a majority to form, making them largely blind to the “miracle of aggregation.”