The Output-Agreement Method Induces Honest Behavior In The Presence Of Social Projection

The output-agreement method has been successfully used to reward agents in a variety of crowd-sourcing settings. This method consists of a simple payment function that randomly matches two agents reported information and rewards agreement. In this letter, we discuss how the output-agreement method might induce honest behavior when there exists social projection, i.e., when each agent believes that his private information is the most popular one amongst his peers.