Research 

How does political knowledge moderate the effect of new information on political attitudes and behavior? In this article, using two key assumptions from John Zaller’s Receive-Accept-Sample model, I reconcile previously disparate findings in the literature and identify conditions under which we should expect low-information voters to exhibit high-levels of opinion change. I revisit Zaller’s characterization of “Floating Voters” —individuals with low-levels of political awareness who are particularly responsive to election specific information—and find similar attitudinal and behavioral effects for these individuals when faced by large informational shocks across two case studies with two different dependent variables. I find that low-knowledge voters were more likely to increase their demand for defense spending after the September 11, 2001 attacks, and that they were more likely to state an intention to vote for Barack Obama after the Lehman Brothers collapse in 2008. These findings suggest that low-information voters change their opinions in an expected and coherent manner when faced with large informational shock. Finally, I show that by explicitly engaging with the assumptions of RAS, questions which are usually reserved for the study of electoral behavior can be studied in non-election contexts. These findings enrich our understanding of the relationship between political knowledge and the dynamics of public opinion and provide theoretically motivated predictions of when we should expect greater opinion change among the least informed.

Imputation of missing racial and ethnic identifiers is essential for studying the political behavior of minority groups. Despite the diversity and growing importance of Asian-Americans in American politics, methods to accurately identify sub-groups within this population perform poorly, particularly for Indian-Americans. Current methods either rely on coarse racial classifications or depend on outdated ethnicity-specific surname lists with declining sensitivity. I address this problem by using data from the Indian Electoral Rolls to construct a comprehensive Asian-Indian surname list that is temporally robust. The underlying premise is straightforward: surname lists derived from the sending country capture the full distribution of immigrant surnames in the receiving country, unlike sample-based methods. In contrast to previous approaches, this dictionary eliminates the need for costly updates to track immigration-driven surname changes.

Racial disparities are widespread throughout the U.S. justice system, in arrests and incarceration. These disparities are typically explained by appealing to racial biases among the police and the judiciary. I present a model in which disparities arise between groups despite unbiased actions on the part of these authorities. If voters can influence the intensity of legal enforcement and the minority population is heterogeneously distributed across counties, then, in the presence of an in-group bias, state-level disparities can arise. State-level variation in judicial selection methods generates a testable prediction from the theory; we should expect the intensity of policing to be positively correlated with minority population size in counties in which judges are appointed compared to counties in which they are elected. Using a county-level panel of arrests between 2000-2014 in the United States, I find support for this hypothesis. A 1% higher share in the black population is associated with a 0.58% increase in the clearance rate of property crimes. I do not find a comparable effect in states with elected judges.

How do individuals know whether their preferences are moderate or extreme? Are there differences in the speed at which individuals learn where their preferences lie relative to each other? This paper presents a simple model of learning in which individuals invest effort to learn where their preferences rank in the population through a random matching process. Under certain assumptions on the coarseness of the signals exchanged in this process, I find that individuals whose latent policy positions are more extreme become more certain of where their preferences lie relative to the population faster than others. The implications of these results are discussed with relation to electoral competition, campaign donations, interest group formation, and as a possible microfoundation for preference intensity.