What Research Participants Learn from Rewards: A Conversational Logic Analysis of Rewarding Reasoning Performance
StatisticsView Usage Statistics
Based on a conversational logic framework (Grice, 1975), we examined the implications of rewarding performance on a reasoning task. We argued that, in the interaction between experimenter and participant, financial incentives signal that a problem requires effort and that its solution is difficult to find. Thus, we hypothesized that reward would lead participants to distrust obvious, but correct, solutions and lower their task-related confidence. Two experiments (n = 33 and n 36) tested these predictions using various versions of the Wason four card selection problem. Consistent with our prediction, rewarded participants were less likely than non-rewarded participants to include an obvious and correct card in their final solutions. At the same time, rewarded participants worked longer on the task, found it more difficult and were less confident in their answers.