ABSTRACT
What people answer in surveys is not necessarily what they would do in reality. In a survey experiment administered to 1650 respondents in the United States, I show participants the profile of a fictional charity operating refugee camps. Participants tend to be more generous toward the charity when asked to make a hypothetical future donation than with an immediate real donation. This phenomenon is known as hypothetical bias. We know very little about the individual-level characteristics associated with hypothetical bias. In this article, I focus specifically on survey respondents who claim they would donate hypothetically but then decline to do so in reality. Since very few studies compare answers to hypothetical and real-life charitable donation requests, I draw on research comparing self-reports to actual actions in the context of normative behaviors such as voting and giving. The same motivations that may lead someone to overreport their generosity or to claim that they voted when they did not can help us explain hypothetical bias. Based on this literature, I hypothesize that more religious individuals, more highly educated individuals, and individuals of color are more likely to display hypothetical bias. As predicted, I find that religiosity, education, and race are all associated with hypothetical bias. Beyond contributing to the theoretical understanding of hypothetical bias, this research can help nonprofit fundraisers allocate resources more effectively.