Voters’ Verdicts

Citizens, Campaigns, and Institutions in State Supreme Court Elections

Chris W. Bonneau, Damon M. Cann

• Publisher: University of Virginia Press
• Year: 2015
• Print ISBN: 9780813937595
• eBook ISBN: 9780813937601
• Topic: Politics & International Relations
• Subtopic: Political History & Theory
• Index: Politics & International Relations

In Voters' Verdicts, Chris Bonneau and Damon Cann address contemporary concerns with judicial elections by investigating factors that influence voters' decisions in the election of state supreme court judges. Bonneau and Cann demonstrate that the move to nonpartisan elections, while it depresses political participation, does little to mute the effects of partisanship and ideology. The authors note the irony that judicial elections, often faulted for politicizing the legal process, historically represented an attempt to correct the lack of accountability in the selection of judges by appointment, since unlike appointive systems, judicial elections are at least transparent.

This comprehensive study rests on a broad evidentiary base that spans numerous states and a variety of electoral systems. Bonneau and Cann use the first national survey of voters in state supreme court elections paired with novel laboratory experiments to evaluate the influence of incumbency and other ballot cues on voters' decisions. Data-rich and analytically rigorous, this provocative volume shows why voters decide to participate in judicial elections and what factors they consider in casting their votes.

A volume in the series Constitutionalism and Democracy

1

INDIVIDUAL-LEVEL FACTORS AND VOTER PARTICIPATION IN STATE SUPREME COURT ELECTIONS

When you go into the ballot box and see a whole list of judges’ names on a ballot, and you don’t know anything about them, it’s a hopeless mess. People just, in many cases, don’t vote at all.

—Sandra Day O’Connor, “Keynote Address by Justice Sandra Day O’Connor”

The conventional wisdom about the nature of judicial elections, espoused by interest groups, the media, and policymakers, includes two claims: first, voters do not know anything about judicial elections or the candidates running in them; second, in many cases voters opt not to vote at all. We address the first point in chapters 2 and 3. In this chapter, we address the second claim and investigate the determinants of voter participation in state supreme court elections focusing on the impact of institutional arrangements.

For decades, scholars of politics have been interested in the determinants of political participation. In order for elections to fulfill their purpose as effective mechanisms by which citizens can participate in the functioning of government, it is imperative that we understand the factors that make citizens more (or less) likely to participate in the political process. Scholars have spilled copious amounts of ink examining the causes of political participation, largely focusing on federal elections (Freedman, Franz, and Goldstein 2004; Henderson and Chatfield 2011; Kam and Palmer 2008; Krasno and Green 2008; Riker and Ordeshook 1968; Rosenstone and Hansen 1993; Sondheimer and Green 2010; Wolfinger and Rosen stone 1980). Only recently have political scientists turned their attention to down-ballot elections (but see Dubois 1979; Bonneau and Hall 2009; Bullock and Dunn 1996; Hall 2007; Schaffner, Streb, and Wright 2001).

In this chapter, we seek to answer a simple yet unanswered question: what factors influence an individual’s decision to participate in a state supreme court election? More specifically, we are interested in the differences between partisan elections and nonpartisan elections. Do partisan elections increase voter participation compared to nonpartisan elections? At least since the time of the Progressives over a hundred years ago, partisan elections have been frequently criticized for depressing the number of candidates, thereby stifling competition, and for making election officials beholden to their own political party at the expense of their constituents (Duncan 1913; Gould 1986; Hofstadter 1955). Regardless of the empirical validity of these charges, the costs of partisan elections may be outweighed by their benefits. Here, we examine one such potential benefit: increased voter participation.

In order to examine the effects of the partisan ballot on voter participation, we first need an office where occupants are selected by both partisan and nonpartisan elections. Fortunately, state supreme courts provide exactly such an environment, allowing us to isolate the effect of ballot type while holding the office constant. Indeed, state supreme courts have three forms of elections by which judges are selected: competitive partisan elections (partisan affiliation of candidates on the ballot), competitive nonpartisan elections (no partisan affiliation of candidates on the ballot), and noncompetitive retention elections (where judges are simply subjected to a yes/no vote on whether they should remain in office). In this chapter, we limit our analysis to the comparison between partisan, quasi-partisan, and nonpartisan elections. We examine retention elections in chapter 6.

The second thing we need is individual-level data on voter participation across different states. Without this, we cannot ascertain the determinants of the behavior of individual voters. While existing studies of voter participation in state supreme court elections have gone a long way toward explaining the variance in voter participation rates (Bonneau and Hall 2009; Dubois 1979; Hall 2007), these studies all utilize aggregate-level data (but see Johnson, Schaefer, and McKnight 1978). Thus, these studies cannot tell us anything about the individual-level variables determining participation in judicial elections. We use the 2010 and 2012 nationwide CCES surveys in order to alleviate this limitation.

Our results indicate that the factors that predict general turnout differ to some extent from those that determine participation in judicial elections. We show that state institutional factors, particularly the partisan format of the election, as well as individual factors such as education, gender, and union membership are strong predictors of participation in judicial elections. In sum, consistent with prior work and contrary to some claims, voter participation in state supreme court elections is affected by voter characteristics, the context of the election, and the rules under which the election occurs.

Partisanship and Voter Participation

The debate over the effects of partisanship on voter participation dates back to the early twentieth century. As mentioned above, the Progressives embarked on a campaign to purge partisanship from elections. As it relates to voter participation, the Progressives argued that “an active and properly motivated citizenry does not need party labels” (Schaffner, Streb, and Wright 2001, 8). The electorate should simply choose the candidates that are best qualified to run the government effectively and efficiently; party labels are unrelated to these characteristics. Thus, elections should be nonpartisan.

However, removing partisan affiliation from the ballot deprives voters of a key piece of meaningful information. It has been well established in the political science literature that one of the best cues for voters to determine which candidates best reflect their ideology and policy positions is the partisan affiliation of the candidate (e.g., Aldrich 1995; Page and Shapiro 1992). That is, if a voter is a Democrat, then any Democratic candidate for an office will better represent the voter’s policy positions than any Republican candidate for an office. As Schaffner, Streb, and Wright (2001, 9) say, “It follows, then, that taking party labels away in nonpartisan elections and thereby raising the costs of information about candidates for voters, nonpartisan elections would make voting more difficult and thereby undermine the potential for popular control.” Due to the removal of an important piece of information from the ballot, voters may lack enough information to feel sufficiently informed to cast a vote. When voters feel uninformed, participation will decrease (Downs 1957; Hall 2007).

Schaffner, Streb, and Wright (2001) find evidence consistent with this. They find lower levels of voter participation when a non partisan ballot is used compared to a partisan ballot (see tables 4 and 6). In judicial elections, Dubois (1979) was the first to find that there was higher participation in partisan elections compared to nonpartisan elections. More recently, Hall (2007) and Bonneau and Hall (2009) document that, on average, there are higher amounts of ballot roll-off (lower amounts of voter participation) in states that utilize nonpartisan elections versus partisan elections. In addition to the aggregate data, Bonneau and Hall (2009) examine two states that recently switched from partisan elections to nonpartisan elections (Arkansas and North Carolina). They find that voter participation decreased significantly in both states after the move to nonpartisan elections. As further evidence of the effects of switching away from partisan elections, Dubois (1979, 875) examined four states that switched from partisan elections to retention elections in the 1948–74 period, and the results “demonstrate the dramatic depressing effect upon turnout when voters are deprived simultaneously of meaningful voting cues and an electoral choice.” Thus, there are good theoretical and empirical reasons to suspect that nonpartisan elections serve to decrease voter participation.

Participation and Judicial Elections

The advantage of examining voter participation in judicial elections is that these elections provide three different types of election in which we can examine voter participation. This source of institutional variation forms the basis for our primary hypotheses.

In some states, judges are selected in partisan elections. These are equivalent to the traditional type of election for most major offices in the United States: each major political party (Democratic and Republican) has a candidate on the ballot, and the partisan affiliations of each are noted on the ballot so voters know, at the time they cast their votes, which candidate represents which party. We hypothesize that voter participation will be highest under this institutional setup. There are two reasons for this. First, as discussed above, this provides voters with an important piece of information that will help guide their vote choice. When voters feel more informed, they participate in higher numbers (Bonneau and Hall 2009; Downs 1957; Hall 2007). Second, and relatedly, voters who vote a straight party ticket will also have their vote registered for judicial candidates as well. Thus, all voters who select the straight party option have their vote recorded for every representative of that party on the ballot; there is no possibility for voters to skip an office. Consequently, even if the voters had no information about the candidates, if they were to vote a straight ticket, their vote would be recorded for all offices (Bonneau and Loepp 2014).

Another method of judicial selection is nonpartisan elections. In these states, multiple candidates run in contested elections; however, the partisan identification of the candidates is not provided on the ballot to the voters. Indeed, it is possible for both candidates to be from one political party in these states. This lack of information leads us to expect that voters will be less likely to participate in these elections compared to partisan elections. Additionally, voters who utilize the straight party ticket option would not have their votes counted for the judicial races, since these races are nonpartisan. Individuals who vote a straight ticket would need to vote for judicial races separately; voters might not be aware of this and thus might accidentally fail to participate in the election even when they intended to do so (Bonneau and Loepp 2014).

Two states, Ohio and Michigan, use a quasi-partisan hybrid of the partisan and nonpartisan systems. In this system, parties are allowed to nominate and endorse candidates as in a partisan system. However, the ballot itself is nonpartisan because candidates’ party affiliations are not shown alongside the candidates’ names (as they are in other partisan elections). We expect voter participation in these races to be higher than it is in traditional nonpartisan elections (since there is party involvement) but lower than in traditional partisan elections (since the partisan affiliation of candidates is not on the general election ballot, and Michigan has the straight-ticket voting option, meaning that voters who utilize this will not have their votes recorded for the state supreme court races).¹

Finally, some states utilize retention elections to keep judges on the bench. These elections are structured more like referenda than they are competitive elections, though unlike referenda incumbents in retention elections nearly always win. Unlike partisan and nonpartisan elections, retention elections do not provide the voters with a choice between candidates. Rather, voters are simply asked whether Judge X should keep his or her job. If a majority of them agree,² then the judge keeps his or her position for another fixed term of office (six to twelve years). Retention races are characterized by a lack of campaigning and high affirmative vote totals for the incumbent (Bonneau and Hall 2009). As such, we omit retention elections from this chapter, leaving participation in retention elections for chapter 5.

The presence or absence of partisan labels in judicial elections is, at its core, an informational problem. From among the host of individual-level variables that predict voter turnout generally, the individual voter characteristics that we expect will have the greatest impact on turnout will be those that are related to voters’ levels of information. For example, we would expect education to influence participation in judicial elections more than home ownership because education has more direct bearing on a person’s level of information than home ownership.

In sum, we hypothesize:

H1.1: A voter is more likely to participate when he or she is provided with the partisan affiliations of the candidates.

H1.2: Individual-level factors related to the level of information of citizens will have larger effects than individual-level factors that are simply demographic in nature.

Data

To better understand the dynamics of voter participation in judicial elections, we utilize data from the 2012 and 2010 Cooperative Congressional Election Study (CCES). The CCES is a national stratified matched sample of voters that is administered through a web-based platform on computers or other web-enabled devices. Cross-validation of the CCES against mail and random-digit dial telephone methods suggests that the CCES provides a nationally representative sample (Ansolabehere and Schaffner 2011). The CCES is segmented into modules of a thousand respondents each, with the first half of the survey composed of common content asked of respondents on all modules. The second half of the survey is devoted to team content created by the respective sponsoring institutions. More information about the CCES can be found in appendix A.

On the 2010 CCES, judicial vote choice questions were asked on three modules of one thousand respondents each. Because we can only use responses from states where there were contested judicial elections, our sample size using the 2010 data is just over n = 1000. The 2012 CCES featured judicial vote choice on the common content, giving us thousands of respondents who lived in states holding contested judicial elections. In 2012, all CCES respondents living in a state holding a contested judicial election were asked about their participation in one randomly assigned state supreme court race in their state. A list of the races analyzed in this chapter can be found in appendix B.

Rather than examining voter turnout in judicial elections, because few, if any, voters come to the polls simply to participate in a judicial election, it is more appropriate to look at the extent to which voters who have already come to the polls to participate in up-ballot races decide to cast a vote in a judicial election. The failure to vote in a down-ballot race after voting for higher offices is known as ballot roll-off. For example, if one hundred people turn out to vote for president and ninety of these people vote in the state supreme court race, ballot roll-off is 10%. Higher levels of ballot roll-off reflect lower levels of voter participation.

We can take an initial look at the data as a simple test of H1.1 by simply showing the proportion of individuals in each election format who voted in a race for higher office (using verified turnout),³ but who did not report voting in a judicial race. For this initial examination, we rely on only the 2012 CCES due to its larger sample size (the sample includes 7,265 individuals who voted for higher offices who could have also chosen to vote in a contested state supreme court election). Figure 1.1 shows the proportion of ballot roll-off by election format (nonpartisan, quasi-partisan, and partisan).

Consistent with the findings of previous studies, our survey results suggest that nonpartisan elections have the highest rates of ballot roll-off of the three election formats we consider, with nearly one-fifth of voters opting to not participate in a judicial election in spite of voting for highe...

Table of contents

• Cover Page
• Title Page
• Copyright
• Acknowledgements
• Introduction: The Landscape of Judicial Elections
• Dedication
• Contents
• 1. Individual-Level Factors and Voter Participation in State Supreme Court Elections
• 2. Vote Choice in State Supreme Court Elections
• 3. Voting Decisions in Partisan and Nonpartisan Elections
• 4. Does One Good Term Deserve Another? Incumbency and Vote Choice
• 5. Voting Yes: Retention Elections
• 6. Conclusion
• Appendixes
• Notes
• References
• Index