Ballots used in the American states differ in the information that they provide about candidates. This article examines the effects of ballot information on voters' decisions whether and how to vote in individual contests, effects on which scholars have not yet provided direct evidence. The study employs experimental manipulations within pre-election and post-election surveys on three contests for seats on a state supreme court. Ballot information on candidates' incumbency status and city of residence had little impact on voters' decisions. In contrast, information on candidates' party affiliations had substantial and interconnected effects on participation and choice. These findings illuminate voter decisionmaking in low-information contests, and they demonstrate that the choice between partisan and nonpartisan ballots in judicial elections is indeed consequential.
One lesson of the recent presidential election-that not all ballots are the same-came as no surprise to political scientists. They have long noted and studied differences in ballots and the effects of these differences on voter behavior. However, one specific characteristic of the ballot-the information it provides about candidates-has received very little systematic study. This is true even though ballot information varies greatly across jurisdictions and has been extensively debated by policymakers and reformers. As a consequence, we lack direct evidence as to the magnitude of the impact (if any) that ballot information has on voters, what forms that impact takes, and what types of ballot information have an impact.
The study reported here is designed to provide such evidence in the context of judicial elections. We use a set of experiments embedded within pre-election and post-election surveys to isolate the effects of providing three types of ballot information that some states offer to voters. Because the experiments focus directly on ballot information and reproduce important elements of an actual voting experience, they will provide the clearest and most extensive evidence yet brought to bear on the effects ballot information has on individual voters' choices.
What we find should help us understand the interplay between electoral mechanisms and voter decision-making better, casting new light on the ways in which voters reach choices in low-information contests-especially what information they use and how much they rely on it. At the same time, the study can contribute to normative discussions of judicial selection systems. We will return to these points in our concluding discussion.
The Issues and Evidence
There are good reasons to think that the content of the information on a ballot could influence voters' decisions. For one thing, there is considerable evidence that other characteristics of the ballot have an impact. Researchers have demonstrated that the structure of ballots (e.g., office block versus party column) and the actual mechanisms of voting (e.g., punch-card versus electronic voting) can affect the participation rates of voters already at the polls (Walker 1966; Darcy and Schneider 1989; Nichols and Strizek 1995). Ballot structures may increase or reduce ticket-splitting (Allen 1906; Rusk 1970). And the order in which candidates' names appear on the ballot may help determine the number of votes they receive (Brook and Upton 1974; Miller and Krosnick 1998; but see Darcy 1998).
Further, information is central to electoral choices. Presumably, voters' plans as they enter the booth are shaped largely by the information they possess. For highly visible and salient contests, voters typically encounter sufficient information to settle on a candidate before entering the voting booth. But for contests enjoying limited visibility and salience-such as nearly all judicial races-they usually enter the polls with a dearth of stored information. In such races, many voters may find their information insufficient to allow a firm choice between candidates. …