Incentives in Experiments with Objective Lotteries

When subjects in an experiment are given multiple decisions, their choices in one decision may be distorted by the choices made in others. An experiment’s payment mechanism is incentive compatible if no such distortions occur. Azrieli et al. (2014) provide two characterizations of incentive compatible mechanisms in a general decisiontheoretic framework in subjects’ choices are represented as Savage-style acts. In particular, paying for one randomly-chosen problem—the Random Problem Selection (RPS) mechanism—is incentive compatible when we assume preferences satisfy event-wise monotonicity, and nothing else. Here, we consider the case where subjects view gambles as objective lotteries. Using completely different proof techniques, we show that the set of incentive compatible mechanisms under the monotonicity assumption is strictly larger than in the acts case. We discuss these new incentive compatible mechanisms in detail.