I'm Sure I'm Right: Identifying the Effects of Confidence in Political Knowledge

• In Event: Knowledge, Misperceptions, and Conspiracies

Abstract

Overconfidence is among the most studied cognitive biases of the past three decades and researchers are increasingly concerned not only with what citizens know about politics, but with what citizens think they know. Past work examining the behavioral consequences of confidence in political knowledge is primarily correlational in nature, identifying associations that are presumed to be causal. We introduce a flexible method of exogenously manipulating confidence in one’s knowledge within the context of a survey, allowing causal identification of the effects of confidence on attitudes and behaviors across a wide variety of knowledge domains. We present survey respondents with a common set of political knowledge questions and provide feedback about the accuracy of their responses. To manipulate confidence, we vary the criteria used to judge responses as correct across experimental conditions: respondents who receive easier versions of knowledge questions are expected to exhibit more overconfidence, as a result of more extensive memory search and explicit negative feedback about the accuracy of their responses. Across three national surveys (N = 7,000) we validate this method and use it to causally test frequently hypothesized relationships between political overconfidence and behavioral outcomes, including belief-updating, information-seeking, and endorsement of political misperceptions and rumors.

Authors

• Brian Guay, Massachusetts Institute of Technology

• Adam J. Berinsky