Abstract

Many attitudes rest on a network of underlying and related beliefs. For example, capital punishment might seem compelling because of the belief that it is a deterrent; it might seem undesirable because it requires the state to judge who may live and die. In this dissertation I address four questions about attitudes and their underlying beliefs. First, I examine whether individuals change their attitudes in response to evidence that disconfirms their underlying beliefs. Conclusions in the existing literature include: people ignore such evidence; people acknowledge it but in a way compatible with their biases; and people polarize their beliefs in response to disconfirming evidence. These conclusions rest on questionable data. In particular, prior research considers beliefs in a circumscribed way, without taking account of the network of information in which they are imbedded. In this study, I provide people with disconfirming evidence for their belief and for the related information that supports the belief. When this is done, attitude change occurs more often than polarization. Second, I describe the cognitive strategies used by individuals to respond to disconfirming evidence. Abelson's theoretical account remains the received and empirically untested view of the reasoning strategies used by individuals confronted with conflicts between their attitudes and new information. The current study empirically tests Abelson's theoretical account and finds it to be incomplete. Third, I examine whether strategy use is content-free. The goal of providing content-free prescriptions for reasoning has a long history. However, those who prescribe formal models of reasoning fail to consider that such models can only be applied by taking content into account. In this light, I consider Gibson's concept of affordances in the physical world and discuss whether a similar phenomenon exists in the informational world. Finally, I examine whether individuals' responses to disconfirming evidence are contingent upon the psychological functions served by their attitudes. I explore whether conclusions about the role of attitude functions in reasoning about other issues apply to reasoning about capital punishment, and whether this role remains important when attitudes are addressed in more than a circumscribed way