Thank you for following the work of the AEI-Brookings Election Reform Project. We’ll continue looking at the issues of election reform at AEI and Brookings. For new work on congressional redistricting, please visit

Viewpoint: If My Election Isn't Broken, Don't Fix It: The Limits of Voter Confidence Measures and Public Opinion on Election Reform
Molly Reynolds, Senior Research Coordinator, AEI-Brookings Election Reform Project

July 1, 2009

The legitimacy of elections ultimately rests upon public acceptance of the results. The tumultuous public reaction upon hearing the "official" results of the Iranian presidential election reminds us that widespread public perceptions of blatant election fraud can threaten even non-democratic regimes. The dead-heat finish and disputed count in Florida that proved decisive in determining the winner of the 2000 U.S. presidential election inflamed the passions of partisans and eventually uncovered many flaws in the administration of American elections, but the vast majority of citizens accepted as legitimate the outcome determined by the 5-4 decision of the Supreme Court in Bush v. Gore. Nonetheless, fixing those flaws was deemed essential to maintaining public confidence in American democracy.

Upon signing the landmark Help America Vote Act (HAVA) in 2002, President George W. Bush proclaimed that "every registered voter deserves to have confidence that the system is fair and elections are honest, that every vote is recorded and that the rules are consistently applied" and that "the legislation I sign today will add to the nation's confidence." Seven years later, policymakers at the federal, state, and local levels are still working to create this fair, accurate, and consistent election system that Bush touted. What can we conclude about the current level of public confidence in American elections? Does the voting experience of citizens shape their confidence in the system? Is public opinion a useful guide to policymakers in considering reforms of the process?

While President Bush laid out in plain language a claim that election reforms, such as HAVA, would produce a public confident in the administration of its elections, the results remain somewhat mixed. On one hand, voters are unambiguously confident that their own votes are counted accurately. In the 2008 Cooperative Congressional Election Study (CCES),[1] 83% of respondents indicated before the election that they believe their own ballots would be counted correctly on Election Day; that number jumped to 90% when respondents were asked retrospectively after the election about the counting of their own votes. More doubt, however, remains in the context of the administration of elections nationwide. In the same survey, only 48% of respondents surveyed prior to the election were either very or fairly confident that ballots "across the country" would be counted accurately. While this number increased to 64% in the post-election survey, the gap in perception between the likelihood of one's own ballot being counted and the probability of everyone else's receiving the same treatment is certainly worth noting.

What, then, might explain this difference specifically and could determine voter confidence more generally? Work by other scholars drawing on data from the same CCES instrument offers a number of suggestions as to what makes people confident in the accurate counting of their own votes. In one analysis, Paul Gronke and James Hicks find that voters' confidence in the accurate counting of their own ballot is strongly related to the quality of the general experience they have when they visit the polls on Election Day, as well as more specifically to their interaction with the poll workers at their precinct.[2] A second preliminary analysis by Paul Herrnson, Richard Niemi, and Kelly Patterson, while focusing more broadly on voters' satisfaction with the voting process, suggests that pre-election expectations also play a key role--and one independent of the actual experience individuals have when they visit the polls--in determining their overall satisfaction with the process.[3] While the analysis does not go as far as to explore the possible origins of the expectations that, in turn, shape voters' assessments of their experience, the authors posit that prior positive experiences could produce higher expectations and thus more positive evaluations of subsequent visits to the polls. Both analyses, then, suggest that voters' assessments of and confidence in the election system are affected by what actually happens when they go to vote. While in most cases, this dynamic likely means that voters who have positive voting experiences--as Herrnson, Niemi, and Patterson find that most individuals do--will continue to do so and will remain confident in the system, it also implies, conversely, that one bad experience on Election Day could drastically alter the picture for that particular voter.

These analyses, then, contribute to a clear story about the connection between personal voting experience and confidence that one's own vote is counted accurately: if one has a good experience on Election Day, one is confident that his or her vote gets counted. This confidence, in turn, may make him or her likely to have higher expectations about future trips to the polls. If he or she expects things to go well during the next election, he or she is more likely to be satisfied with the process, which feeds back into to high levels of confidence about his or her own vote. This story also, moreover, contributes to an explanation of why voters are less confident in the ability of the system as a whole to count votes accurately. Voters' confidence in the accurate counting of their own ballots is clearly intimately tied to their own experiences at the polls. Asking voters about the overall election administration system, however, requires them to think more broadly and generally and to make assessments based on what little information they have about the conduct of elections in other parts of the country. That information is almost certainly going to be negative (memories, for example, of Florida in 2000 or Ohio in 2004), as, then, the subsequent judgments are likely to be-just as a majority of Americans continually give their own Congressman high approval ratings while disapproving of the institution as whole, or rate their own personal financial situation as stable while simultaneously believing the national economy is severely struggling.

Though most voters are clearly pleased with their own experiences at the polls, then, they also recognize that the system as a whole is imperfect and that there remains room for additional improvement. While much of the public debate over what form additional reform should take immediately after HAVA was dominated by security concerns about electronic voting machines, the focus in more recent years has shifted to concerns about other aspects of the voting experience. Can voters be required to show identification at the polls? Yes, according to the Supreme Court (PDF) in 2007. If presented with convenience voting options-like voting in person before Election Day, voting by mail, or having choice over one's polling location-will they take advantage of them? Statistics from 2008 certainly suggest as such; an estimated 30 percent of voters cast their ballots before Election Day in 2008.

In the coming months and years, more states are likely to explore implementing these types of policies, and the federal government may wade back into the debate in the near future. What specific proposals are they likely to entertain? A look at areas of elite consensus provides some suggestions. Newspaper editorial boards, for example, have come out overwhelmingly against voter ID laws; perhaps most notable, even boards in states like Georgia and Florida that subsequently adopted such laws opposed them in their run-up to their passage.[4] Editorial boards have also embraced early voting and voting by mail as ways to make the process more convenient for busy individuals; they have also tended to endorse Election Day registration (EDR), whereby voters can register at the same time they cast their ballot at the polling place. Reform advocacy groups have long touted similar proposals, promoting EDR, largely rejecting calls for an ID requirement, supporting automatic registration for individuals upon their eighteenth birthdays, and either advocating for, or calling for more research into, the efficacy of voting by mail and vote centers.[5]

Advocacy groups and policymakers may be continuing to shout for reform, but a look at public opinion paints a rather different picture. These reform proposals elites embrace? The public either does not agree, or simply does not express a strong preference either way. In the case of the former, for example, 70 percent of respondents in a national survey support requiring voters to show identification at the polls, while only 17 percent oppose such a proposal. Sixty percent of voters would prefer to continue to vote at their local precinct-based polling place rather than visit a vote center, a much-touted arrangement whereby voters have the option of voting at any one of several large polling sites (located in places like municipal buildings and shopping malls) near where they work or shop instead of being required to go to the smaller site near their home. A plurality of respondents "strongly opposes" (47 percent) making voting mandatory, as it is in Australia and a number of other countries.

The mass public may oppose policy elites on some election reform issues, but on others, opinion is split to the degree that no consensus emerges. Forty-three percent of respondents supported automatic registration for eighteen year-olds, while 44 percent opposed it. The breakdown was identical when respondents were asked about Election Day Registration. When asked about whether they would favor "loosen[ing] constraints on when Election Day has to be held"-whether it be moving it to a weekend or spreading voting out over a several day period, as is done with current early, in-person systems-respondents distributed themselves fairly evenly across six categories, with 33 percent generally favoring the idea, 30 percent generally opposing it, and 36 percent feeling either neutral or lacking an opinion altogether.

What are we to make of this gap between elite and popular opinion, then, and what does it mean for the future prospects of various election reform efforts? Many election reform opponents would be tempted to seize the public opposition and indifference towards these proposals--particularly when coupled with relatively high levels of confidence in and satisfaction with the system--as evidence that elites are making much ado about nothing and that significant additional policy change is unnecessary. A closer look at some possible determinants of the public's opinions on these proposals, however, suggests otherwise. Most voters--as the Herrnson, Niemi, and Patterson data show--both expect their voting experiences to go smoothly and are subsequently generally pleased with the ways things go, making them relatively unlikely to come out in support of significant changes to the system. For these voters, if their own system isn't broken, why fix it? As we know, however, the system is broken for some voters; an analysis of the overall CCES sample suggests that four to five million individuals nationwide wanted to vote in 2008 but did not because of a registration problem or because they did not receive an absentee ballot.[6] Even many of these same respondents who express high levels of faith in the accurate counting of their own ballot recognize this fact, as they have markedly lower levels of confidence in the system as a whole.

While an "if it isn't broken, don't fix it" story likely accounts for some of this gap between mass and elite opinion, some of the explanation may also lie with the inherent shortcomings of a survey's ability to accurately measure opinions. Election reform issues are not especially salient for the average voter, particularly since most of them have very good experiences when they go to the polls. Thus, few are likely to have well thought-out and informed opinions on the topic, and when asked for their thoughts on reform proposals, are likely to simply guess or offer an off-the-cuff answer. To test this hypothesis, we devised two survey experiments to see if opinions were so malleable that they could be manipulated simply by changing the wording of the question. Respondents were asked first about their opinion on laws that require voters to show identification in order to vote, and indeed, they were slightly more likely to support ID laws if they were told that "some people think that voters should have to show [ID] in order to prevent voter fraud" than if they were told other people "think that such a requirement would be unfair since some people don't have an acceptable ID and would have trouble getting one because of the cost or documents required." The effects of framing the question differently, moreover, were even greater when respondents were asked about whether they favored a paper-based voting system or an electronic, ATM-like device. When told that "some people think that voters should make their choices directly on a computer screen and have the computer count the results so that voting can be as easy as using at ATM," 249 respondents favored an electronic system and 142 supported a paper one. Conversely, when told that "others think that voters should make their choices on paper ballots and then run them through a scanner because computer systems might be prone to fraud," 203 supported the electronic option versus 202 favoring a paper-based system. The framing of the question had an obvious impact on the responses.[7] If, then, we can move public opinion on election reform around simply by changing the wording of the questions we ask, to what extent should we rely on it as a guide for what reform strategies to pursue? While few would advocate completely ignoring mass opinion in making policy in any area, the probable existence of more non-opinions than opinions here should at least give policymakers pause as they move forward.

Indeed, at the launch of the AEI-Brookings Election Reform Project in 2006, then-Senator Barack Obama captured several of these dynamics of public opinion on potential changes to our election system. "Election reform," he said, "is one of these issues where America has a tendency to go from shock to trance. We're shocked right after an election when the news reports horror stories about disenfranchisement and intimidation and suppression; there's a public outcry; there's a flurry of legislation; but then the politics gets tough and the problems aren't solved and pretty soon everyone forgets about it until the next electoral crisis." Having avoided such widespread shocks in recent years, the nation remains in a trance of sorts and, as a result, any efforts to enact the types of reform proposals being touted by policy elites will collide with a neutral-at-best, resistant-at-worst public. While the expression of the public's opinions is problematic enough that it should not completely deter reformers from pursuing their goals, it does, nonetheless, add one more hurdle to an already uphill battle.

Molly Reynolds can be reached at

Viewpoint is an occasional feature analyzing various election reform issues.

[1]All public opinion data discussed in this piece comes from the 2008 Cooperative Congressional Election Study; more information on the design of the survey can be found here. In short, 33,000 respondents participated in the entire survey project. Half of the items addressed to each respondent were addressed to all respondents, while the other half were given to that individual respondent and approximately 1,000 other respondents. The question batteries aimed at the smaller groups within the overall sample were developed by teams at individual institutions around the country, with some institutional teams, including ours, collaborating to produce a larger number of respondents for individual items. We thank Costas Panagopoulos of Fordham University, Steve Ansolabehere of Harvard University, and, in particular, Paul Gronke and James Hicks at Reed College/the Early Voting Information Center and the Pew Center on the States for their assistance in formulating various items on the survey instrument and sharing their data on those items.
[2]See Paul Gronke and James Hicks, "Re-Examining Voter Confidence as a Metric for Election Performance," Paper presented at the annual meeting of the Midwest Political Science Association, 2 April 2009.
[3]Early results of this analysis were presented at the 2009 CCES Conference, May 28-30, 2009; a more detailed version, entitled "Expectations and Satisfaction with the Voting Process in the 2008 U.S. Presidential Election," will be presented at the meeting of International Political Science Association, Santiago, Chile, July 12-16 July, 2009.
[4]This discussion of editorial board positions on election reform proposals relies on data collected via a Nexis search of editorials in major U.S. newspapers since January 2005.
[5]The websites of ten major reform advocacy groups were surveyed to ascertain their positions on various issues. The groups investigated were: the Brennan Center for Justice, Common Cause, Demos, Commission on Federal Election Reform (the Carter-Baker Commission), the Lawyers' Committee for Civil Rights Under Law, the Century Foundation, Project Vote, the Election Center, the League of Women Voters, and FairVote.
[6]See Ian Urbina, "Hurdles to Voting Persisted in 2008," New York Times, 10 March 2009.
[7]Tim Ryan of the University of Michigan (formerly of the Election Reform Project) devised these survey experiments and performed the analysis of the results discussed here. More detailed results of the analysis are on file with the author.

Featured Resources
Convened to provide guidance to the state of Utah in the areas of election law, the Commission made recommendations regarding voter registration. Joining a number of other groups in proposing automatic voter registration, the Commission adopted all of its recommendations unanimously.
Focusing on the four major provisions of the National Voter Registration Act, Project Vote’s report reviews implementation of voter registration programs at the state level.
The opinion in the case of League of Women Voters of Indiana, Inc. v. Todd Rokita, Indiana Secretary of State.
Rather than submit a provisional ballot, some unregistered voters are able to both register and vote on Election Day. Same Day Registration has been enacted in nine states and the District of Columbia. This memo explores this policy’s effect on turnout.
The Research Database on the U.S. Voting System and Voting Technology provides access to empirical and analytical research about voting and elections to inform evidence-based reforms.
Research Projects
Center for Democracy & Election Management was established at the School of Public Affairs at American University in 2002. Their broader goal is to pave the way for and strengthen democracy through improved electoral performance.
Election Law @ Moritz, run through Moritz College of Law at the Ohio State University, contains both explanation and commentary on a wealth of election reform issues from a legal perspective. provides daily news updates on election reform issues, as well as deeper analysis of selected topics, including recent reports on voter registration, recount procedures, and the progress in implementing the Help America Vote Act since 2002.
Dēmos is a non-partisan public policy research and advocacy organization founded in 2000. A multi-issue national organization, Dēmoscombines research, policy development, and advocacy to influence public debates and catalyze change.
Project Vote is a national nonpartisan, nonprofit 501(c)(3) that works to empower, educate, and mobilize low-income, minority, youth, and other marginalized and under-represented voters.
The American Enterprise Institute for Public Policy Research
The Brookings Institution
© Copyright 2016, AEI
and The Brookings Institution