Title: Payment Rules and their Effect on Effort and Truthful Reporting on Micro-Task Markets
Abstract: In the domain of crowdsourcing there exist online marketplaces where workers are paid by employers in exchange for their labor. In cases where workers are intended to report private information to the employer there arise challenges achieving truthful reporting and effort elicitation. One approach to facilitate truthful reporting is the use of payment rules which couple the worker’s beliefs about a ground truth to the occurrence of future events. In this talk we will present two experiments which address the effects of different payment rules and risk preferences on effort elicitation and truthful reporting. The first experiment compares two approaches of how to assess subjects’ risk preferences. The second experiment uses different incentive compatible and incentive incompatible payment rules and compares their effect on effort and truthfulness of the workers’ reports. These experiments result in the following two contributions: Firstly, we show that two common approaches of defining subjects’ risk preferences can result in conflicting outcomes. Secondly, we demonstrate that while incentive incompatible payment rules can be exploited by subjects when repeated over multiple rounds, they can also lead to a higher elicited effort compared to incentive compatible payment rules.