To what degree is voter confidence in election procedures driven by satisfaction with the outcome of an election, as opposed to trust in government or objective features of the polling place, such as voting technology? Using approximately 30 national surveys over the past decade, we find a consistent relationship between voting for the winner and confidence in election administration. This confidence varies as a function of question wording and electoral context. Respondents are more confident in the quality of the vote count locally than nationally. They are responsive to electoral results at the state and national levels in forming their judgements. And, rather than being influenced by different types of voting technology, respondents lose confidence by virtue of change itself.Keywords: Voter confidence, legitimacy, public opinion, winner effect, election administration (2008), ruled that Indiana's interest in assuring the public of the integrity of the election process partially justified the passage of a photo identification requirement in order to vote.
1As the improvement of voter confidence has been promoted as a goal of election reform, and as scholars have begun probing voter confidence through survey research, questions have arisen about whether expressions of confidence in these surveys are a product of voters' direct experience of voting, or whether they are a product of something else, such as the causally prior belief about the trustworthiness of government more generally. More troubling for democracy, there is also the possibility that expressions of confidence in how ballots are counted are no more than reflections of whether one's preferred candidate won or lost the election: if my preferred candidate won, then I think the election was conducted fairly. To the extent this is true, it not only calls into question the usefulness of confidence questions for guiding policy makers, but it also suggests that contentious disputes over election outcomes will be inescapable in an increasingly polarized society, no matter what administrative reforms are put in place.1 Gronke and Hicks (2009) contains a comprehensive review of recent studies that examine voter confidence as a dependent variable, as well as a summary of recent policy debates that center around the concept of voter confidence.
3Our purpose in this paper is to explore this last possibility, by combining data from a number of survey research projects. Specifically, we focus on asking: to what degree is voter confidence driven by a respondent's satisfaction with the outcome of an election, as opposed to more general trust in government or objective features of the polling place, such as voting technology? While all the data will be observational, we are able to exploit several features of the data that help support a causal interpretation of our results, such as the dynamic nature of the data and the unpredictability of certain election results.We find there is indeed a consistent relationship between voting for the winning candidate A...