CURIOSITY, CONFIDENCE and DISCREPANCY DETECTION

Curiosity is a defining characteristic of human intelligence. The results of people’s curiosity include artistic creativity, technological progress, and scientific discovery. Despite the importance of curiosity, there have, until recently, been relatively few advances in our scientific understanding of the nature of curiosity itself, of its theoretical characterization, of the conditions that promote or thwart it, and of its direct consequences. Current theories of curiosity embrace two different and seemingly incompatible ideas. On the one hand, curiosity is strongly goal directed and reward oriented; it beckons individuals toward the tantalizing goal of the almost-known solution. On the other hand, curiosity is goal averse; it lures individuals away from the beaten path, spurning nearer rewards, and tempting individuals to explore novel territory. Both of these views about curiosity are compelling, yet they also seem irreconcilable. This project tests an alternative characterization, which is compatible with both views, that curiosity is an attentional state that highlights and is selectively sensitive to discrepancies. A person in such a state would be more likely to detect errors (including their own errors), and to embrace new information. Hence, curiosity puts people in a perfect position to learn. This research tests the hypothesis that curiosity is associated with an attentional state that amplifies error detection and promotes learning. In addition to the planned research, there are complementary activities related to providing rigorous research opportunities for trainees and outreach through school partnerships.A series of planned experiments tests the hypothesis that when people are curious they are particularly sensitive to errors and that their curiosity-based detection of errors can be used to allow them to correct those errors. Two planned experiments are directed at determining neural correlates of error-related curiosity. Other planned experiments test how curiosity-based detection and correction of errors may be leveraged to enhance learning and memory. In past research, curiosity seemed to provide a marker that people both 'knew' that they were wrong, and that they were eager to find the correct answer. A series of experiments, based on the finding that people are especially curious about their high confidence errors, aim to both isolate those errors and help overcome them. The plan is to compare the effects of different methods of feedback on learning, specifically for factual questions about which people express both high knowledge and high curiosity (but on which they are wrong). One type of planned feedback is direct correction, in which people are given the correct answer and told to learn it. There are also plans to test both scaffolded feedback, in which people are given hints until they come up with the answers themselves, and self-sought feedback, in which people are given the opportunity to look up the answers online. It is hypothesized that these latter, more active, kinds of feedback will produce larger and more lasting effects on learning. Furthermore, it is hypothesized that because of people's high curiosity, they will enjoy these highly interactive methods of obtaining feedback, and the process may stimulate further curiosity and eagerness to learn. Overall, the experiments in this project test how people's curiosity about their own errors may improve their learning and memory in a proactive, lasting, and enjoyable manner.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.