Evidence of a detrimental effect of extrinsic incentives on breaking a mental set.

Abstract College students were either rewarded ($1.50) or not for solving a series of 10 water-jar problems. The first 9 of these were designed to establish a set for a relatively complicated, three-jar solution (B − A − 2C). The 10th problem was a set breaker which required that subjects discover a simple (A-C) solution. The mental activities needed to produce a solution to the two types of problems (Problems 1 to 9 and Problem 10) were, therefore, different and one interest was in seeing whether reward would be detrimental for performance on both types of problems. It was not. Poorer reward group performance was obtained only on Problem 10. This result was not due to between-group differences in mathematical ability, and explanations in terms of differences in time taken to check answers or concern over quality of performance were considered and discounted. Measures of intrinsic interest (subject evaluations of the task and a measure of willingness to volunteer for future research) failed to support the belief that interest differences account for performance differences.