There is a persistent gender gap in motivations to run for political office. While exposure to role models is widely believed to increase women’s political ambition, there is little field experimental evidence on whether exposure to female politicians in realistic settings can increase political ambition. We conducted a field experiment in which a sample of 612 female students was randomly assigned to receive emails inviting them to an event that included career workshops with female politicians, or no email. The treatment increased interest in the ongoing national election campaign, but, against expectations, did not have any positive effect on political ambition. Our results suggest that female politicians who discuss their experience bluntly, instead of following a motivational script, may fail to motivate other women to pursue a political career. These results highlight the need for more research into the type of events and messages that bring more women into politics.

Muslim Americans constitute one of the United States’ most vulnerable minority groups, facing frequent discrimination from both the public and the government. Despite this vulnerability, few studies evaluate interventions for reducing prejudice against Muslim Americans. Building from an insightful literature on the sources of prejudice against Muslim Americans, this paper tests whether attitudes can be improved with information countering misperceptions of the community as particularly foreign, threatening, and disloyal to the United States. The experimental treatment modestly improved attitudes, including among some subgroups predisposed to prejudice against Muslim Americans. However, the treatment struggled to change policy views, and it demonstrated some vulnerability to social desirability bias and priming on terrorism threats. The findings suggest that information campaigns addressing misperceptions can help to reduce prejudice on the margins, but primarily in less politicized contexts.

Results of an audit study conducted during the 2016 election cycle demonstrate that bias toward Latinos observed during the 2012 election has persisted. In addition to replicating previous results, we show that Arab/Muslim Americans face an even greater barrier to communicating with local election officials, but we find no evidence of bias toward blacks. An innovation of our design allows us to measure whether e-mails were opened by recipients, which we argue provides a direct test of implicit discrimination. We find evidence of implicit bias toward Arab/Muslim senders only.

Campaign experiments often report positive effects on voter turnout. But do these effects endure at subsequent elections? Existing studies provide mixed evidence on downstream effects, and the rate at which initial mobilisation effects decay. This paper contributes to existing research by presenting a pre-registered analysis of downstream effects in a unique experimental setting. I test whether effects from a UK partisan experiment in a low turnout election in May 2017 persisted at the high turnout general election a month later. The findings show that in this short space of time, the original turnout effects virtually disappeared, suggesting that downstream effects resulting from campaign experiments can be quickly subsumed by the high saliency of subsequent elections.

How should education be structured to most effectively increase civic outcomes such as political knowledge and democratic values? We present results from a field experiment in which we compare the effects of deliberative education and traditional teacher-centered education. The study is the largest field experiment on deliberative education to date and involved more than 1,200 students in 59 classrooms. We test the effects on four forms of civic competence: political knowledge, political interest, democratic values, and political discussion. In contrast to previous research, we find little evidence that deliberative education significantly increases civic competence.

Political scientists rely heavily on survey research to gain insights into public attitudes and behaviors. Over the past decade, survey data collection has moved away from personal face-to-face and telephone interviewing towards a model of computer-assisted self-interviewing. A hallmark of many online surveys is the prominent display of the survey’s sponsor, most often an academic institution, in the initial consent form and/or on the survey website itself. It is an open question whether these displays of academic survey sponsorship could increase total survey error. We measure the extent to which sponsorship (by a university or marketing firm) affects data quality, including satisficing behavior, demand characteristics, and socially desirable responding. In addition, we examine whether sponsor effects vary depending on the participant’s experience with online surveys. Overall, we find no evidence that response quality is affected by survey sponsor or by past survey experience.

In many situations, incentives exist to acquire knowledge and make correct political decisions. We conduct an experiment that contributes to a small but growing literature on incentives and political knowledge, testing the effect of certain and uncertain incentives on knowledge. Our experiment builds on the basic theoretical point that acquiring and using information is costly, and incentives for accurate answers will lead respondents to expend greater effort on the task and be more likely to answer knowledge questions correctly. We test the effect of certain and uncertain incentives and find that both increase effort and accuracy relative to the control condition of no incentives for accuracy. Holding constant the expected benefit of knowledge, we do not observe behavioral differences associated with the probability of earning an incentive for knowledge accuracy. These results suggest that measures of subject performance in knowledge tasks are contingent on the incentives they face. Therefore, to ensure the validity of experimental tasks and the related behavioral measures, we need to ensure a correspondence between the context we are trying to learn about and our experimental design.

Does the process used to pass a law affect the way citizens evaluate the outcome? In a series of experiments, I manipulate the way in which a law is passed – ballot initiative or the legislative process – to test the effect of process on citizens’ evaluations of policy outcomes. The results show that people view the ballot initiative process as fairer than the legislative process, but that process has a negligible effect on outcome evaluations.