Prediction Errors in Two Domains

Abstract

This dissertation investigates how social interactions fail, and why people are so often unaware of the things they do to cause those failures. Two of the main reasons why social interactions fail is that people fail to communicate effectively and people fail to allocate resources optimally. The three papers in this dissertation focus on various aspects of these two topics. Paper 1 demonstrates that people think it will be better to have an extraordinary experience even when they have to interact with people who had an ordinary experience instead. These predictions turn out to be wrong because people who had an extraordinary experience were unexpectedly excluded from a subsequent conversation. Paper 2 shows that people appreciate the joy but underestimate the difficulty of communicating about novel experiences. Speakers in these studies predicted that listeners would enjoy hearing about a novel experience more than a familiar one, but listeners were actually much happier hearing about a familiar experience because novel stories were confusing. Paper 3 investigates how people often try to please their interaction partners by using fair procedures to allocate important and valuable resources. Across several studies, allocators used fair and unfair procedures to distribute money, and although allocators predicted that receivers would care about fairness, these predictions were incorrect. In sum, this dissertation investigates a series of prediction errors that people make when they try to communicate and allocate resources. The dissertation then goes on to explore the reasons behind these errors and the consequences for people’s well-being and their social interactions.