Publications

Many states allow some form of non-traditional precinct voting, from voting early in-person to various forms of mail-assisted balloting. Research on how these voting reforms impact voter participation has produced a wide range of mixed findings. We review this literature and describe how the research designs and measures used by past studies may have biased their results. We then offer a new theoretical approach for testing the relationship between convenience voting reforms and voter turnout that addresses some of the pitfalls in previous measurements of reforms. We also identify best practices for how to implement our new measures using a difference-in-difference research strategy that is robust to many confounding factors and describe a potential sustainable database for undertaking this research over time.

Absentee ballots preserve the ability of voters to participate in elections in the event they are unable to vote in person. Texas has used absentee voting for over 95 years, however the law requires an excuse must be given for why an accommodation is necessary. Using a survey experiment from a statewide survey, I breakdown how the threat of contracting a disease is likely to alter voter behavior. The results show that population density and recognition of a public health threat provide additional marginal effects beyond difference of opinion by party. When all voters face a public health crisis, like the coronavirus pandemic, voters are more likely to consider voter convenience in their interest for fair and legitimate elections.

Are voters as polarized as political leaders when it comes to their preferences about how to cast their ballots in November 2020 and their policy positions on how elections should be run in light of the COVID-19 outbreak? Prior research has shown little party divide on voting by mail, with nearly equal percentages of voters in both parties choosing to vote this way where it is an option. Has a divide opened up this year in how voters aligned with the Democratic and Republican parties prefer to cast a ballot?

We address these questions with two nationally diverse, online surveys fielded from April 8-10 and June 11-13, of 5,612 and 5,818 eligible voters respectively, with an embedded experiment providing treated respondents with scientific projections about the COVID-19outbreak. We find a nearly ten-percentage point difference between Democrats and Republicans in their preference for voting by mail in April which has doubled in size to nearly twenty-percentage points during the pandemic. We also find that support for national legislation requiring states to offer no-excuse absentee ballots has emerged as an increasingly polarized issue.

News media coverage on Election Day often focuses on longlines at polling places. News outlets are incentivized to cover long lines, as they make for dramatic television–despite long waits being rare. Coverage of long lines potentially deters turnout due to public perception of long wait times, reduces confidence in election administration, and undermines trust in government institutions. However, we know little about the potential deleterious effects of emphasizing longlines. Improving understanding of Election Day coverage has important implications for scholars, election officials, and journalists.

This paper has two parts: First, a systematic content analysis of television news coverage of voting on Election Day: we analyze coverage of Election Day in the 2016 and 2018 elections on six national television news networks (ABC, CBS, NBC, Fox News, CNN, MSNBC) plus a sample of local network news. Second, we will conduct a survey experiment in which each subject is randomly assigned to a television news story about polling place lines with a different news frame: a baseline condition with a news story unrelated to voting and treatment conditions varying the video imagery, chryons (bottom of screen text characterizing story), and reporter voiceover. We will measure attitudes about vote intention, confidence in elections, expected wait time, blame attribution for long lines, and credibility and fairness of the coverage

Voters who use the straight party voting option (SPVO) are more likely than others to roll off when voting for nonpartisan offices and ballot questions. Previous research has theorized that this effect is due to voter error, as individuals fail to understand that they must still complete nonpartisan questions after selecting the straight party option. Using cast vote records, we find the SPVO leads both to voter error in some individuals and satisficing in others. About half of voters using the SPVO who engage in nonpartisan roll-off leave all nonpartisan elected offices blank. At the same time, among voters who vote for at least one nonpartisan elected office, individuals using the SPVO are still more likely than other voters to engage in nonpartisan roll-off. Survey data confirm both of these patterns. Among voters using the SPVO, those with college education are more likely to state that they intentionally rolled off rather than were confused about their vote while those without college education are more likely to be unsure about who they voted for. We also find that the effect of the straight party option on nonpartisan roll-off increases with voter fatigue and decreases with campaign spending, supporting the explanation of satisficing. Our results highlight the unintended effects of ballot design in influencing voter behavior.

Large-scale ballot and survey data hold the potential to uncover the prevalence of swing voters and strong partisans in the electorate. However, existing approaches either employ exploratory analyses that fail to fully leverage the information available in high-dimensional data, or impose a one-dimensional spatial voting model. I derive a clustering algorithm which better captures the probabilistic way in which theories of political behavior conceptualize the swing voter. Building from the canonical finite mixture model, I tailor the model to vote data, for example by allowing uncontested races. I apply this algorithm to actual ballots in the Florida 2000 election and a multi-state survey in 2018. In Palm Beach County, I find that up to 60 percent of voters were straight ticket voters; in the 2018 survey, even higher. The remaining groups of the electorate were likely to cross the party line and split their ticket, but not monolithically: swing voters were more likely to swing for state and local candidates and popular incumbents.

Many democracies changed their electoral systems since 1990s either to permit or restrict the access of small parties into parliament. In some others, governments initiated an electoral system change, yet failed to enact the reform as in the examples of New Zealand, the Netherlands, and Poland. These failed electoral reforms receive scant attention in the literature, save for a few case studies. In this paper, I model electoral reforms as driven by strategic calculations of the ruling-party whose preferences about alternative electoral systems are shaped by the dynamics of the party competition. I specifically focus on the impact of small parties on the competition between the largest two parties and develop different scenarios of permissive and restrictive reforms. This novel account expects that the ruling party initiates an electoral reform depending on whether small or new parties draw votes from its vote base or from that of its main competitor in the election. For the success of reforms, I examine the role of institutions, in particular the level of institutional protection that electoral systems have. I test the hypotheses by using an original dataset of cross-national electoral reform attempts in 32 parliamentary democracies between 1945and 2015.The findings of the study support the main hypotheses. I find that ruling parties are more likely to initiate an electoral reform restricting small party access when small parties draw votes from its vote base, but an opposite one when small parties draw votes from its main competitor in the election.

A model is developed to forecast monthly county level voter registration totals by party for Florida, Iowa, Maryland, Oregon and Wyoming. A prediction model of change in voter registration can help election administrators prepare for periods of high workload. Another application is to aid the adjustment of voter registration figures with an eye towards removing deadwood registrants. The aim of doing so is to boost voter registration data’s efficacy in predicting phenomena such as postcensal population growth, election outcomes and voter turnout. Descriptive statistics on the concentration of drops and negative net changes in total registrants are presented to discuss how much information such data can yield to help estimate deadwood in conjunction with prediction models utilizing demographic and other factors