Surprise, Surprise: The Role of Surprising Numerical Feedback in Belief Change

Abstract

What percentage of U.S. residents is incarcerated?If you now learned the amount and it surprised you, would it be more memorable than if it were not surprising?Our past research documented conceptual changes related to policy issues when one receives a single, critical number.In the present study, Experiment 1 uses a Numerically-Driven Inferencing (NDI) paradigm method in which participants estimated policyrelevant quantities, learned the true quantities, and rated their surprise regarding that feedback.When asked to recall the quantities either eight or 84 days post-feedback, participants improved the most over their original estimates on items that surprised them the most.In Experiment 2, we found that a measure of prospective surprise ("shock"; Ranney, Cheng, Nelson, & Garcia de Osuna, 2001)-derived from an interval in which participants believed the number fell, and participants' confidence that the number fell in that intervalreliably predicted retrospective surprise ratings like those in Experiment 1.We conclude that surprise is a rather stable construct about which people have considerable metacognition.Future work in this area may suggest how leaders, voters, and consumers can best employ their emotional responses to numbers and enhance cognitive strategies that help shape effective policies.