Incentive effects: The case of belief elicitation from individuals in groups

► We elicit probabilistic beliefs from experienced observers about choices in a 2x2 asymmetric matching pennies game. ► Observers are either paid based on a quadratic scoring rule or just the same flat fee per round in the treatments. ► Beliefs are elicited a second time in each round where subjects can revise their forecasts after observing the beliefs of all other members of their group. ► The lack of incentives for accurate beliefs has a substantial negative impact on the accuracy of initial as well as updated individual and group mean beliefs.