Dissertation: How Political Communication Shapes Economic Opinion and Action


A critical question in contemporary U.S. politics is why support for redistributive policies is not higher, given the economic struggles of many American families. My dissertation argues that individual-level support for redistributive policies reflects not only citizens’ objective economic conditions, but also their subjective economic status and the ways in which redistributive policies are politicized in electoral campaigns. Using public opinion surveys, survey experiments, and observational data, I demonstrate that campaign efforts to politicize economic problems like unemployment and immobility are most likely to succeed at changing minds and encouraging political participation in places experiencing economic growth. In areas with high levels of unemployment and inequality, efforts to politicize job loss actually discourage political engagement. The implication of this research is that individuals are less likely to perceive proposed redistributive policy interventions to be viable when they perceive their own economic prospects to be limited.

Politicizing Unemployment (Job Market Paper)
Campaign advertising politicizes issues, representing economic and social problems as reasons to engage in political action. But efforts to politicize economic issues are conditional on economic reality, which shapes voter perceptions of proposed policy interventions. I use observational data to examine the relationship between job-focused political advertising, local labor market conditions, and voter turnout. Since campaigns cannot control which counties within a media market view economic issue-focused ads, individuals in very different economic contexts are exposed to the same level of aggregate issue-focused advertising. Using both difference-in-difference and cross-sectional analyses, I show that when campaigns broadcast more job-focused ads, voter turnout increases in areas with low levels of unemployment and decreases in areas with high levels of unemployment. These results suggest that issue-focused ads are an imperfect means of encouraging political participation: individuals who might benefit most from policy interventions are demobilized by ads that focus on those policies.
Visualizing Inequality: How Graphical Frames Shape Public Opinion is forthcoming in Research and Politics. Replication materials are available here. The supplemental appendix is available here.
Graphical representations of political issues and economic trends are an increasingly popular means of conveying information to the public. However, graphs have the potential to shape public opinion by visually emphasizing or downplaying the information they convey. I randomly assign subjects to view graphs that represent the same underlying information but that differ in relative emphasis: one is consistent with a textual account of rising inequality, while the other de-emphasizes the same information by increasing the scale of the Y-axis. My results indicate the graphical frames provide powerful contextual cues: for Republicans and conservatives, exposure to the de-emphasizing graph results in a 40% decrease in expressed support for intervention against inequality relative to Republicans and conservatives in the control condition, despite the fact that both groups read the same textual information. My findings reveal how an increasingly important and unexamined form of political communication affects public opinion, also suggesting promising avenues for future research.
Opportunity is Relative: Localized Economic Mobility Information and Redistributive Preferences
Information about economic mobility is politically salient and relevant to ongoing debates about equality of opportunity. To test how local and national economic mobility information shapes public preferences, I conduct a nationally representative survey experiment using local mobility estimates from Chetty et al (2014). In the experiment, I randomly assign respondents to receive no information, information about average mobility across U.S. commuting zones, or both national and local mobility estimates. Both treatments change attitudes toward a range of opportunity-increasing redistributive policies but local economic context moderates observed effects: those in low mobility areas who receive the national treatment become more supportive of government intervention, while those in low mobility areas who receive the national and local treatment become less supportive of intervention. I argue that individuals’ relative perceptions of how information applies to their own local context help explain these counterintuitive findings. These results suggest that campaign appeals that focus on low mobility rates might undermine support for intervention, depending on the geographic specificity of the appeal.

Political Communication, Public Opinion, and Political Behavior

Spillover Effects in Political Advertising: Evidence from Judicial Elections (with Thomas Gray) is under review.
We develop a simple theory of information spillovers within political advertising: individuals use information conveyed by candidate-specific ads to evaluate other candidates seeking different offices. We test the theory by linking state court of last resort election returns with a complete set of gubernatorial, congressional, and state supreme court political advertisements in 2010. We show that up-ballot advertisements about crime have spillover effects on judicial race outcomes. Specifically, fearful references to crime in non-judicial campaign advertisements decrease judicial incumbent vote share, while enthusiastic references to crime increase incumbent vote share. We estimate that cumulative spillover effects from crime ads in gubernatorial and congressional advertising campaigns in 2010 decreased incumbent justices’ average expected vote share, affecting the outcome of at least one state supreme court election in 2010. We show that ads affect vote choice by providing information to voters, rather than by priming local crime rates. We also conduct two placebo tests to validate our results. Overall, these findings provide evidence of ads’ political effectiveness, while our research design proposes a conceptual framework for understanding advertising effects in context.
Extremism and Political Compromise (with Andrew Clarke) is in progress.
Compromise is central to democratic deliberation and legislative policymaking, but opinion scholars have only recently begun to assess how the public thinks about compromise in a systematic way. We define rational compromise as an individual’s willingness to accept policies that are different from her most preferred policy. Using a simple spatial model, we show how individuals’ willingness to accept policy alternatives depends upon the policy status quo and the relative extremism of the individual’s initial preference. To test the model, we develop original, interactive survey instrumentation that allows us to directly assess individuals’ willingness to compromise across different political issues. Our results show that extremists act rationally, accepting a larger range of alternative policies than moderates.
Measuring Total Survey Error: A Dynamic Factor Model Approach (coauthored with Jee-Kwang Park) is under review.
Opinion polls are often inaccurate, but the relative effects of sampling and non-sampling sources of error on poll accuracy remain difficult to discern. However, elections provide a unique opportunity to measure the composition of survey error because underlying, true public opinion is revealed on election day. Using dynamic factor analysis, we examine the survey error structure of 2012 presidential pre-election polls conducted by eleven major polling organizations and estimate the size of sampling and non-sampling errors. We find that a large portion of total survey error is attributable to non-sampling error: across the eleven pre-election polls we examine, bias accounts for about two-thirds of survey error in terms of mean square error. We also show that sources of non-sampling error have a large effect on the variance of polling estimates, which does not decrease with increasing sample size. These results confirm and extend the theoretical arguments of the Total Survey Error (TSE) paradigm.
Grocery Shopping for America: Terrorism, National Identity, and Consumer Behavior (with Sonal Pandya and Raj Venkatesan) was presented at IPES 2014 and is in preparation for journal submission.
Does national identity have a causal effect on economic behavior? Survey experiments in International Relations provide causal estimates of preference with a high degree of internal validity. These results’ value as theoretical microfoundations however depends on the correlation between survey responses and behavior, and the consistency of preferences across contexts and over time. Supermarket purchases are a frequent, consistent, and universal type of economic behavior. We analyze weekly supermarket sales following the 9/11 terrorist attacks to show an increased market share for American-sounding brands, brands consumers perceive to be American by virtue of nationality branding cues. The attacks increased the salience of national identity to American consumers. Findings are robust to controls for price, product availability, and seasonality. Experimental results show national identity correlates with American-sounding brand preferences more than product price and other subject demographic characteristics. 9/11’s effect on brand preference does not persist in 2014. These findings suggest supermarket sales are a rare behavioral correlate of international political preferences.
National Solidarity in Wartime: How Military Casualties Shape Consumer Behavior (with Sonal Pandya and Raj Venkatesan) was presented at APSA 2015 and IPES 2015 and is in preparation for journal submission.
A growing body of survey-based evidence suggests social identities drive foreign policy preferences but to build microfoundations of IR theory two puzzles remain: (a) social identity has to drive actual behavior, and (b) variation in identity’s salience over time and across contexts must be readily observed. We analyze US consumer behavior following war casualties, an exogenous, quasi-randomly distributed shock that makes national identity more salient. In our analysis of household panel data, including 90,000 U.S. households, and retail scanner data, including over 70 million individual consumer purchases, we match Iraq and Afghanistan war casualties during 2002-2008 to their US hometowns to assess local shifts in behavior. Purchases of American-sounding supermarket brands increase ten fold in communities following the death of a local solider. Households that included armed forces employees show especially large increases. These findings demonstrate a casual effect of national identity on economic behavior and suggest that casualties influence attitudes about war independent of media and partisan priming.

Public Policy and Political Behavior

Detecting Diffusion: Assessing the Performance of Event History Analysis for Studying the Adoption of Policy Innovations (with William BerryCraig Volden, and Scott Liebertz) was presented at APSA 2015.
We simulate policy adoptions based on diffusion and non-diffusion mechanisms in order to assess the performance of standard event history analysis (EHA) models.  We evaluate the extent to which EHA models yield false positives for variables that do not play a driving role in the simulations and false negatives for driving variables.  We also assess whether the marginal effect estimates for driving variables accurately match true marginal effects in the data generating processes.  We find relatively low false positive rates, but identify substantial problems with the ability of EHA to uncover significant and accurate effect estimates for true driving variables.  For low-probability policy adoptions, EHA performs particularly poorly.  For higher probability adoptions, EHA has only about a 50% success rate in reporting accurate and statistically significant estimates.
Rights, Responsibilities, and the Ideology of Mental Health (coauthored with Lynn M. Sanders and Nicole Pankiewicz) was presented at APSA 2011, and is in preparation for journal submission.
In the field of mental health politics and policy, the rights revolution has produced a fundamental contradiction. Liberals, or left-leaning civil libertarians, are more ambivalent proponents of expanded mental health services than we might otherwise expect, for fear of constructing an administrative apparatus or service delivery system that could undermine consumer rights. Meanwhile, those on the mental health right, who worry that too many civil liberties mean disordered and dangerous individuals on the streets, are impatient with services restrictions, whether they originate with state budget-cutters or managed care companies. To test this hypothesis, we analyze public opinion data from the last two decades about mental health service provision.
Socializer or Signal: How Agency Accreditation Affects Organizational Culture (coauthored with Manny Teodoro) was published in Public Administration Review.
Agency accreditation has grown steadily as a management strategy in recent decades. Accreditation is meant to help professionalize public administration work by requiring an agency to adopt policies and practices that are sanctioned by an external organization. Advocates claim that accreditation facilitates the diffusion of best practices and builds a culture of professionalism in an agency. Accreditation clearly leads agencies to adopt formal policies. This article identifies two ways in which accreditation might affect organizational culture: (1) by socializing employees, and (2) by signaling the agency’s priorities to employees. Analyzing attitudinal data from officers in six American police departments, this study finds no association between accreditation and officers’ own values, but finds that accreditation is strongly correlated with officers’ perceptions of their agencies’ priorities
Assessing Professionalism: Street-Level Attitudes and Agency Accreditation” (coauthored with Manny Teodoro) was published in State and Local Government Review. Replication materials are available here.
Professional organizations now accredit state and local agencies across several fields. This article investigates the attitudes of street-level bureaucrats toward their agencies and the tasks required of them in accredited and nonaccredited agencies.  If accreditation changes bureaucratic attitudes in ways that build a sense of mission or shape functional preferences, it could lead to more effective public service delivery.  The empirical subjects of this study are American municipal police departments and the Commission on Accreditation for Law Enforcement Agencies (CALEA).  Analysis of quasi-experimental data from a survey of officers suggests that accreditation helps build a sense of mission at the street level but finds little evidence of an effect on functional preferences.