The effect of election administration on voter confidence: a local matter?

Article: “The Effect of Election Administration on Voter Confidence: A Local Matter?” Author: Lonna Rae Atkeson, Kyle L. Saunders Issue: October 2003 Journal: PS: Political Science & Politics This journal is published by the American Political Science Association. All rights reserved. APSA is posting this article for public view on its website. APSA journals are fully accessible to APSA members and institutional subscribers. To view the table of contents or abstracts from this or any of APSA’s journals, please go to the website of our publisher Cambridge University Press (http://journals.cambridge.org). This article may only be used for personal, non-commercial, or limited classroom use. For permissions for all other uses of this article should be directed to Cambridge University Press at permissions@cup.org . The Effect of Election Administration on Voter Confidence: A Local Matter? he 2000 presidential election was a wake-up Telection call to elected leaders, public officials, and scholars. The electoral fiasco—most prominent in Florida, but also taking place in states like New Mexico and Ohio—revealed many deficiencies in voting equipment ~Caltech0MIT Voting Technology Project 2001!. In addition to faulty equipment, registration mix-ups and problems with absentee ballots led to the loss of as many as six million votes ~Caltech0MIT Voting Technology Project 2001!. Confusing ballots, like the butterfly ballot in Florida’s Dade County, were found to have led voters to vote incorrectly ~Wand et al. 2001!. While these problems have, no doubt, existed for a long time, the closeness of the 2000 presidential race and the fact that the number of lost votes had the power to change the election outcome have brought election administration questions to the forefront of policy making. In response, the Help America Vote Act ~HAVA! was passed overwhelmingly by both houses of Congress on October 29, 2002, and was the first ever federal comprehensive law on electoral administration, as election administration has always been the purview of the by 50 states and literally thousands of local adLonna Rae Atkeson, ministrators. HAVA proUniversity of New Mexico vided for $3.9 billion to upgrade older election Kyle L. Saunders, equipment, especially Colorado State University punch-card systems, and established the Election Assistance Commission and minimum election administration standards for the states and local governments who are mostly responsible for administering elections. Despite the implementation of HAVA requirements in the presidential election of 2004, including the change in many states to electronic touchscreen voting, the problems continued. This time the focus was on Ohio, but no doubt the problems were seen elsewhere. First, there were problems with new electronic machines; there were anecdotal reports of machines over reporting votes, for example. Precincts in many urban areas lacked enough voting equipment, which lead to long voter lines and suggested a bias against minority voters. Exit polls in key battleground states showed large discrepancies with the actual vote outcomes, suggesting additional questions about the accuracy and fairness of the election process. Finally, Robert F. Kennedy, Jr. ~2006! published an article in Rolling Stone Magazine that raised questions about the 2004 election PSOnline www.apsanet.org outcome and that ultimately asked: “Was the 2004 Election Stolen?” Though election scholars could not find evidence of systematic fraud, the media coverage and the apparent problems continue to raise doubts about our election system. Voter confidence in our election system is crucial because elections are the link between citizens and their elected officials. In a representative democracy it is the ballot box that allows voters to send their elected leaders mandates for policies and hold them accountable. If voters do not have confidence that their vote is counted correctly then the most fundamental aspect of representative democracy, the direct election of its leaders, is in doubt and a crisis in democracy may be evident.1 From a normative perspective, even though there will always be election winners and losers, voters should still have the utmost confidence in their electoral system. Theoretical Background Voter confidence in elections is a very specific measure of government support in a functional democracy. It broadly lies with a class of measures that examine a broad array of political support including political efficacy and trust in government. These latter measures have been the focus of political science activity since the 1950s, when regime support proved to be quite high in America ~American National Election Studies 2007!. Over time scholars have remained very attentive to the changes in these measures and have noted a rather large erosion in confidence in government in the United States and in Europe ~Dalton 1999!. These changes are disconcerting to many scholars because measurements of regime support capture the commitment to the government and governing structure. Theoretically, greater levels of government trust and efficacy lead to democratic stability and economic security, while lower levels have the potential to destabilize a government and create economic insecurity. One potential problem with studies and measures of confidence in government is that they have focused on broad and generally diffuse measures of regime support. Oft-asked questions about “people in the government” or “people running government” pay very careful attention to the affective component of the evaluation, but little regard for the object or experience of the evaluation ~e.g., Hill 1981!, yet both components make up an attitude. Moreover, these diffuse measures of support DOI: 10.1017/S1049096507071041 655 are likely the product of more specific levels of support that are acquired through interaction and experience with the political system. We argue that studying specific areas of support offer the most likely means for assessing voters’ most frequent experience with the political system, the electoral process, and potentially provide ways to create more positive political experiences that ultimately increase diffuse levels of support and hence strengthen the political system. Indeed, some have argued that trust in the electoral process is pivotal to a democratic society ~Rahn, Brehm, and Carlson 1999; Price and Romantan 2004!. This is because trust at this level builds confidence in other government institutions and individual attachment to the political system. By focusing on global support measures that emphasize the institutions of government, we may be missing primary linkages in understanding citizen connection to the governing process, and those may be key in assessing and understanding the changing nature of voter satisfaction, more broadly speaking, with government. In addition, voter confidence focuses on the procedures of a democracy, as opposed to its institutions, providing us with an alternative referent for assessing the health of our democracy. Recent scholarship also suggests that voter confidence is also related to voter turnout ~Alvarez, Hall, and Llewellyn 2006! and therefore understanding its dynamics may be helpful in building an efficacious and active citizenry. Therefore, we focus our attention on a specific, but vastly important and primary support attitude: whether a voter believes her vote will actually be counted as intended. In our study, the phrase “voter confidence” refers to this support measure. In the aftermath of the 2000 and 2004 elections, such questions are at the administrative forefront of the world’s leading democracy and are thus extremely important questions for scholarly consideration. Data The data we use come from a mixed-mode ~Internet and mail! survey conducted just after the 2006 midterm election. We collected data from respondents in two congressional districts, New Mexico’s First and Colorado’s Seventh Congressional District. We chose these districts, in part, due to the presence of highly competitive congressional contests; ergo, these districts are more likely to see increased voter interest in electoral outcomes and activity as well as the potential for greater problems at the polls. Perhaps more important for the question of voter confidence, both Colorado and New Mexico have recently undergone myriad reforms in their voting laws in response to interest-group pressure to create fair, accurate, and voter-verifiable election administration systems. New Mexico, for example, is the first state to move from a predominantly electronic voting system to one that mandated optical scan bubble paper ballots statewide, with the intent of providing a paper trail so that elections could be audited for accuracy. Further, New Mexico passed legislation to implement a statewide 2% audit, beginning in 2007, to ensure the accuracy and fairness of election outcomes. Meanwhile, Colorado has been the frontrunner in the implementation of many innovative changes, including vote centers, and recent changes to the law mandate a paper trail to ensure voter integrity. In the 2006 cycle, some Colorado voters had the option of choosing either touchscreen systems with voter-verifiable paper rolls or optical scan ballots. In addition, both states have been early adopters of early voting as well as no excuse absentee voting, resulting in many voters choosing to cast their ballots prior to Election Day. In New Mexico about one in five voters took advantage of early voting and about one in five voted absentee, leaving just over half of voters voting on Election Day. In Colorado about half of voters voted absentee and a little over 656 one in 10 voted early with the remainder voting on Election Day. Thus, our sample provides us with interesting variation in voter interaction with the electoral process to assist us in evaluating voter confidence. Survey Design Just before Election Day, we sent 4,050 letters to a random sample of registered voters in both congressional districts requesting their participation in our Election Administration Survey.2 The letter explained our study and its importance, the respondent’s unique position within it, and provided a URL 3 to a web page through which they could enter the survey. The web page presented respondents with FAQ and IRB policies. The letter also explained that respondents could request a mail survey and a return self-addressed stamped envelope by contacting us via a toll free number or by calling our offices. Sample registered voters who did not respond were recontacted three times with a postcard reminding them of the study, the URL, their ability to request a mail survey, and their identification number for the survey. The response rate for the sample was about 15% and was calculated as the number of surveys returned to us, either through web submission or returned mail, divided by the total number of survey respondents who were eligible.4 Survey questions asked respondents about their election experience ~voter confidence, voting problems, method of voting, experience with poll workers, voter satisfaction!, faith in the election process ~including the ability of the machines to provide paper audits!, attitudes toward fraud, voter access, and voter identification as well as other political attitudes and behaviors including evaluations of the president, the congressional candidates, and their local and state election administrators. We also asked several questions related to the congressional race ~vote choice, political activity, etc.! and a variety of demographics.5 Method We focus our attention here, however, on explaining a single measure of voter confidence. We asked respondents: “How confident are you that your vote in the November 2006 election will be counted as you intended?” They could respond “not at all confident,” “not too confident,” “somewhat confident,” or “very confident.” Our primary independent variables focus on three sets of conceptual factors likely related to voter confidence. The first set of factors is related to the voting experience itself, which in some sense is the objective experience the voter has with the voting process ~also see Hall, Monson, and Patterson 2007!. This essentially is the “local” factor and focuses on external attributions in understanding voter confidence. When voters have problems voting,6 for example, because the ballot is confusing,7 or too long,8 or poll workers are unhelpful 9 they are likely to feel less confident that their vote will be counted. And, likewise, the more enjoyable and positive their vote experience the more likely they will feel that their vote will be counted.10 We hypothesize that as the quality of a voter’s first-hand experience with the voting process increases their level of voter confidence increases. Another part of the direct experience is the choice a voter made in how to execute their vote. In New Mexico and Colorado, voters can choose to vote absentee, early, or on Election Day. Voters voting early or absentee are further removed from the election process and may feel less confident that their ballot is likely to be counted. Voters engaging in absentee voting, for example, may feel that their ballot is less likely to be counted because they may believe that these ballots only get counted if PS October 2007 the race is close. A recent national study found Table 1 that absentee voters had significantly less voter Frequency Comparison of Voter Confidence: National confidence, a finding which supports this hypothesis ~Alvarez, Hall, and Llewellyn 2006!. We inversus Local Samples clude both early and absentee dummies in our Local (NM & CO): model, making Election-Day voting the category How confident are you Pew: How confident of reference. that your vote in the are you that your Our second set of conceptual factors relate to November 2006 election vote will be accurately the attitudes that voters bring to the process. The will be counted as counted in the first of these is the voter’s attitude regarding the you intended? upcoming election? machine he used to cast his ballot and whether it offers a verifiable record of his vote. Recall that Very confident 42 58 in New Mexico early and Election-Day voters Somewhat confident 42 29 used identical optical scan bubble paper ballots, Not Too confident 10 9 while in Colorado, most early and Election-Day Not at all Confident 4 3 voters used a touchscreen ballot and some ColoDK/NS 2 1 rado voters had the choice of an optical scan Total N 835 1,503 bubble paper ballot or a touchscreen one. Therefore, we asked voters on a Likert-type scale how strongly they agree or disagree with the following: “The bubble paper ballot @or the touch Results screen ballot# method provides for a paper receipt that can validate the election results.” We then matched voters’ attitudes Table 1 shows the frequency response to our post-election toward the machines to their vote method, creating a scale invoter confidence question compared to a question posed to a dicating how confident they are in the technology they used.11 national survey of registered voters prior to the elections of 2006 We also asked voters about their attitudes toward their immeby the Pew Research Center ~Pew 2006!. Notice that our sample diate election official.12 We argue that the more confidence that is less confident than Pew’s national sample. The national sama voter has in the job her local election official is doing, the ple shows that fully three in five registered voters were very more likely she is to feel confident that her vote is counted. We confident that their vote would be accurately counted, while only asked, “We are interested in whether you strongly approve, aptwo in five felt that way in our sample. While this may suggest prove, disapprove, or strongly disapprove of how your county important differences in pre- versus post-election environments, election official has handled her job.” it may also suggest that local context matters a great deal in The last set of attitudes we focus on is the perceptual lens structuring voter confidence.15 While national data provide an that voters bring to voting through their party identification. important overall look at voter confidence, localized studies may Prior research shows that partisanship plays an important role in be necessary to consider and control for the myriad differences structuring attitudes, including government trust ~Brewer and between election context in voter rules ~e.g., absentee voter Sigelman 2002; Bowler and Donovan 2002; Cook and Gronke rules, registration rules!, election administration differences ~e.g., 2005; Bullock III, Hood III, and Clark 2005; Alvarez, Hall, and availability of vote centers, the voting machines used, history of Llewellyn 2006!, and we suspect that it does here as well. The voting problems in the area0state, competence of election offiproblems in voter administration since 2000 and allegations of cials and poll workers!, and the context of the race ~e.g., compartisan politics, whether in Florida with former Secretary of petitive, media coverage, negativity of campaign!. State Katherine Harris or in Ohio with former Secretary of State Table 2 presents the results of our ordered probit model of J. Kenneth Blackwell, appear to favor GOP political outcomes voter confidence. It also presents a simple symbolic synopsis of over Democratic ones. Therefore, we expect party identification a summary of the comparative effects of our independent varito structure perceptions of the political process with Democrats ables by expressing the change in probability of being in the having less voter confidence than Republicans. normatively important “very confident” category by varying Finally, we also considered a wide variety of demographic each independent variable from its minimum to its maximum controls including gender, age, education ~high school or less, after setting all of our variables to their medians.16 This table some college, college graduate, advanced degree!, income, the reports the direction of the relationship and conveys the comrespondent’s state ~Colorado coded 0, New Mexico coded 1! to parative intensity of change across coefficients as well as the control any differences across election contexts, and race ~nonsignificance levels of the coefficients.17 White coded 0, White coded 1!.13 It is important to note that To begin, our findings demonstrate substantial evidence that self-identified non-White voters in our sample were largely Hispanic.14 voters’ direct experience with the voting process influences their Previous research suggests that political resources, including voter confidence. The more helpful the poll workers and the education and income, increase political efficacy and trust; more a voter enjoyed her voting method, the more confident she therefore our expectation is that they would have a similar posiwas that her vote counted. A more confusing ballot, however, tive effect on voter confidence. Previous research also suggests lowered her voter confidence. Interestingly, we find that not that African Americans are likely to be less trusting in governcasting a ballot on Election Day, but instead voting absentee or ment ~Abramson 1983; Brewer and Sigelman 2002! and several early, results in less voter confidence, especially for absentee studies show that Blacks have significantly less voter confivoting. This last finding is extremely important because increasdence than Whites ~Alvarez, Hall, and Llewellyn 2006; Pew ingly states are providing voters with these alternative means of 2006; Bullock III, Hood III, and Clark 2005!. Therefore we exvoting, yet doing so actually reduces voter confidence in the pect a positive relationship between our race variable and voter process. Such a reduction in confidence is possibly due to the confidence. There are no a priori reasons to hypothesize that disconnection between the voter and Election-Day activities. either gender or age influences voter confidence, but we include When people vote absentee, for example, they may be unsure them as part of our standard model. whether their ballot arrives in time to be counted or they may PSOnline www.apsanet.org 657 Table 2 Ordered Probit Model of Voter Confidence b (Standard Error) Voting experience Voting problems Poll workers helpful Confusing ballot Long ballot Enjoyed voting Method Voted absentee Voted early Voter attitudes Voting Method produces verifiable results County Election Officer Job Evaluation Party Identification Demographics Age Gender (female) Education Race (White) Income State (New Mexico) Cut 1 Cut 2 Cut 3 Chi Square N Model Summary −.202 (.129) .355** (.084) −.281** (.060) −.066 (.043) .199** (.054) −.392** (.118) −.199† (.119) .115* (.054) .259** (.039) .097** (.022) .0001 (.003) −.030 (.090) .073 (.047) .114 (.116) .035** (.014) .131 (.123) NS +++ −−− NS +++ −− −† ++ +++ ++ NS NS NS NS ++ NS .915 (.489) 1.82 (.489) 3.37 (.499) 140.95** 672 Source: Election Administration Study, Atkeson and Saunders (2006) Note: † p < .1, * p < .05, ** p < .01 Notes: Dependent Variable measured on a scale from 1 (not at all confident) to 4 (very confident). OLS regression analysis provided very similar comparative effects. Direction of relationship denoted by type of sign; number of signs denotes comparative change in probability from option 3 (somewhat confident) to option 4 (very confident) when varying each independent variable from its minimum to its maximum and holding all other variables at their medians. The presence of three symbols (pluses or minuses) connotes a change of probability over .30, two signs connote a change of probability between .15 and .30, and one sign connotes a change of probability between 0 and .15. The modeled probability with all variables at their medians is .57 for the very confident category. 658 be uncertain as to whether they filled out the form incorrectly, possibly invalidating their ballot. Early voters likewise may feel more likely that their ballot will be lost or destroyed when the machines are shut off and turned on over the course of days and eventually moved to new locations on Election Day. Such separation from the final moment of the process, when voters can physically observe their ballot being cast on the day it will be counted, increases doubts about the likelihood that their voting effort will be included in determining the election outcome. We also find support for our argument that voter attitudes are important to voter confidence. When voters use a voting machine that they agree produces verifiable results, they are more confident in the election process. Likewise, when they have a positive evaluation of their local county election official, they are also more confident in the election process. The perceptual lens of party is also very important. The stronger the identification with the Republican Party the greater the voter confidence. The last few years of election drama that ultimately has favored Republicans has left a clear mark on voter confidence. This last finding is particularly troubling because the election process should not be seen as a partisan activity. The demographic variables show some expected and unexpected effects. Income is positively related to voter confidence as expected, but education is not. Gender, age, and the state dummy have no relationship to voter confidence, also as expected. Our race variable counter-intuitively shows no relationship with voter confidence. Because race in this study overwhelmingly represents Hispanics and not Blacks, this finding may represent a key difference between minority groups. Blacks’ long history of being denied their civil and voting rights likely plays a role. Interestingly, then, there is no evidence that the Latino voter experience is markedly different from the Anglo experience. Discussion From a public policy perspective voter confidence is a key element of a working democracy and scholastic input into how to make the voting process a better experience for the voter is an important consideration for county clerks and secretary of states across the nation. It is their job to present a process of administration that reduces voter anxiety and increases voter confidence. Our results, therefore, may be helpful to policy makers and administrators and so we make the following broad recommendations. First, local election administrators must work to produce a positive voter experience with as much central guidance as possible. Poll workers must be well trained so that they appear competent, nonpartisan, and helpful to the voter. These direct contacts influence voter confidence and poll-worker training is a direct function of the time, energy, and effort put forth by the local administration officials. Effort in this regard will no doubt reduce conflict between voters and poll workers and increase voter confidence. Ballots must be designed PS October 2007 efficiently and must be unambiguous. Confusing ballots force voters to make mistakes, which reduces their voter confidence. For example, in the New Mexico survey our open-ended questions on why voters rated their overall voting experience fair or poor revealed that the bubbles on the ballot were too small and too difficult to color in for many voters. Larger fonts could ease this problem and potentially make for a better voter experience. And, more generally, we find that when voters use a method of voting they enjoy, they are more confident in their voting experience. Thus, allowing voters multiple voting choices may be key to greater confidence. Again, in New Mexico, the new bubble paper ballots were not well liked by many people, but those were the only machines available to voters. In Colorado, on the other hand, voters in some counties could choose to use a bubble paper ballot or a touchscreen machine. Such options could allow for greater voter choice and produce greater voter confidence. Second, voter attitudes are also important; using machines that produce verifiable results increases voter confidence. Again, this suggests that offering voters a choice of a variety of machines that produce verifiable results may be helpful in producing greater voter confidence. The county election officer is also a key figure. In our study, nearly two in five ~37%! voters could not evaluate their local election administrator, even though this is an elected position. We believe that a more visible role for the local administrator is one factor to greater voter confidence. The county official needs to appear competent, non-partisan, and helpful. Overt or perceived partisanship can reduce voter confidence, as shown by the party identification variable, and therefore the local administrator needs to work hard to be seen as helping all her constituents in the election process. In New Mexico, for example, on Election Day several heavily Republi- can precincts ran out of ballots. While such problems may accidentally happen, the administrator must work hard to prevent them because they undermine her integrity, potentially increase disapproval with her job performance, and hence decrease voter confidence. Voter education, through public service announcements, would assist in connecting voters to their vote administrator and help to create a more positive voter experience and consequently increased voter confidence. Third, our results suggest it is important to look more closely at why early and absentee voting produces less confidence. Many states are increasingly affording these options to their voters, yet our results suggest such options may be problematic for voter confidence. Therefore, at the moment we cannot recommend policies that expand voter options. However, it is unclear as to what underlying mechanism produces this difference and therefore we are very cautious in interpreting the implications of these findings. In conclusion, we call upon more scholarly interest in voter confidence and voter satisfaction questions. The process by which we elect our leaders is at least as important as the trust we place in them once they take office. We show that citizen confidence in the election system is dependent on procedural consistency and perceived fairness and accountability. However, it is also affected by exogenous events in the political context. Without these legitimating forces, democracy, and faith in that democracy, suffers. Many of our conclusions assist in directing efforts toward remedying the problems voters face when they cast their ballots and ultimately toward improving voter confidence. Yet, there is still much work to be done and questions to be answered. Ours is only a first and very small step in this rich and important area of research. Notes * Results were first presented at “The Future of Election Reform and Ethics in the States,” hosted by Kent State University, Department of Political Science, Columbus, Ohio, January 16–17, 2007, and the following paper was presented at the Midwest Political Science Association’s Annual Meeting, Chicago, Illinois, April 12–15, 2007. Data were collected by monies generously provided by the University of New Mexico’s Research Allocation Committee. We’d like to thank Luciana Zilberman, Lisa Bryant, Alex Adams, David Magleby, and the Center for the Study of Elections and Democracy at Brigham Young University for their assistance with this project. Of course, any errors are our own. 1. Though more recently, scholars have also asked if such changes in government support represent a maturation of the public that expresses a healthy but critical electorate ~Norris 1999!. 2. New Mexico Secretary of State Rebecca Vigil-Giron and Colorado Secretary of State Gigi Dennis were kind enough to provide us with the voter registration files updated through the last day of registration activities in 2006 for the congressional districts. 3. The URLs were votenewmexico.unm.edu or votecolorado.unm.edu 4. Due to the poor quality of both states’ voter registration files, over 22% of our sample was unreachable. 5. For an executive summary of our findings, a more in-depth discussion of our sample’s representativeness, and a frequency report of our questions, please visit http:00vote2006.unm.edu. 6. Reported problems include voter name absent from voter list, voter had to vote provisionally, voter had a difficult time finding their polling place, someone else voted under the voters name, lack of proper voter identification, and late or never arriving absentee ballots. 7. We asked, “How confusing did you find your ballot?” Answer choices included: very confusing, somewhat confusing, not too confusing, or not confusing at all. 8. We asked, “How long did you wait in line at your polling place in minutes?” 9. The question was, “How helpful were the poll workers at your voting location.” Answer choices included: very helpful, somewhat helpful, not too helpful, or not helpful at all. 10. We asked respondents to strongly agree, agree, neither agree nor disagree, disagree, or strongly disagree with the following, “I enjoyed voting with the method I used.” 11. Absentee voters were given the median score to prevent listwise deletion in the regression equation. 12. In both constituencies, this is the voters’ county clerk. 13. Summary statistics of all of the variables used in this study can be found in Appendix A. 14. Hispanics represent 12.3%, Blacks 2.1%, Asians 1.5% and American Indians 1%. In New Mexico, Hispanics represent 19% of the sample. 15. There is no difference between Colorado and New Mexico attitudes on this variable. 16. We used the software program Clarify for this task ~Tomz, Wittenberg, and King 2000!. 17. It should be noted that the initial model predicts the probability of being in the “very confident” response option as being 57%, the “somewhat confident” option as being at 39%, the “not too confident” option at 4%, and the “not at all confident” option at barely above zero; these are the baseline probabilities from which the probability estimates in Table 4 vary. References Abramson, Paul R. 1983. Political Attitudes in America. San Francisco: Freeman. Alvarez, R. Michael, Thad E. Hall, and Morgan Llewellyn. 2006. “Are Americans Confident Their Ballots are Counted?” Typescript. Caltech University. PSOnline www.apsanet.org American National Election Studies. 2007. Available at: www. electionstudies.org0nesguide0gd-index.htm#5. Bowler, Shaun, and Todd Donovan. 2002. “Democracy, Institutions, and Attitudes about Citizen Influence on Government.” British Journal of Political Science 32 ~2!: 371–90. 659 Brewer, Paul R., and Lee Sigelman. 2002. “Trust in Government: Personal Ties that Bind?” Social Science Quarterly 83 ~2!: 624–31. Bullock, Charles III, M. V. Hood III, and Richard Clark. 2005. “Punch Cards, Jim Crow, and Al Gore: Explaining Voter Trust in the Electoral System in Georgia, 2000.” State Politics and Policy Quarterly 5 ~3!: 283–94. Caltech0MIT Voting Technology Project. 2001. “Voting: What is, What could Be.” Available at: http:00vote.caltech.edu. Cook, Timothy E., and Paul Gronke. 2005. “The Skeptical American: Revisiting the Meanings of Trust in Government and Confidence in Institutions.” Journal of Politics 67 ~3!: 784–803. Dalton, Russell. 1999. “Political Support in Advanced Industrial Democracies.” In Critical Citizens: Global Support for Democratic Government, ed. Pippa Norris. Oxford: Oxford University Press, 57–77. Hall, Thad E., Quin Monson and Kelly Patterson. 2007. “Poll Workers and the Vitality of Democracy.” PS: Political Science and Politics 40 ~October!: 647–54. Hill, David B. 1981. “Attitude Generalization and the Measurement of Trust in American Leadership.” Political Behavior 3: 257–70. Kennedy, Robert F., Jr. 2006. “Was the 2004 Election Stolen?” Rolling Stone, June 15. Available at: www.rollingstone.com0news0story0 104323340was_the_2004_election_stolen. Norris, Pippa. 1999. “The Growth of Critical Citizens and its Consequences.” In Critical Citizens: Global Support for Democratic Government, ed. Pippa Norris. Oxford: Oxford University Press. Pew Research Center for the People and the Press. 2006. “Democrats Hold Enthusiasm, Engagement Advantage November Turnout May Be High.” News Release, Wednesday, October 11. Available at: http:00peoplepress.org0reports0. Price, Vincent, and Anca Romantan. 2004. “Confidence in Situations Before, During and After Indecision 2000. Journal of Politics 66 ~3!: 939–56. Rahn, Wendy M, John Brehm, and Neil Carlson. 1999. “National Elections as Institutions for Generating Social Capital.” In Civic Engagement in American Democracies, eds. Theda Skocpol and Morris P. Fiornia. Washington, D.C.: Brookings Institution Press. Tomz, Michael, Jason Wittenberg, and Gary King. 2000. “CLARIFY: Software for Interpreting and Presenting Statistical Results.” Version 1.2.2. Cambridge: Harvard University. Available at: http:00gking.harvard.edu. Wand, Jonathan N., Kenneth W. Shotts, Jasjet S. Sekhon, Walter R. Mebane Jr., Michael C. Herron, and Henry E. Brady. 2001. “The Butterfly Did It: The Aberrant Vote for Buchanan in Palm Beach County, Florida.” American Political Science Review 95 ~4!: 793–810. Appendix A Model Variable Descriptive Statistics Variable Median Mean Min Max Voter confidence (DV) 3 3.25 1 4 Early voting (dummy, early = 1) Absentee voting (dummy, absentee = 1) 0 0 0.24 0.41 0 0 1 1 1 1 56 3 7 3 0.76 0.54 54.94 2.73 7.52 3.80 0 0 18 1 1 1 1 1 91 4 16 7 Ever had problems at the polls (yes = 1) How confusing was your ballot (not at all to very) R thought ballot was too long (strongly disagree to strongly agree) R enjoyed method of voting (strongly disagree to strongly agree) R has a positive opinion of county election official (strongly disapprove to strongly approve) Agreed voting method used produced verifiable results 0 1 2 4 3 .18 1.51 2.33 3.75 2.80 0 1 1 1 1 1 4 5 5 5 4 3.88 1 5 State (dummy, New Mexico = 1) 1 0.54 0 1 Race (dummy, White =1) Gender (dummy, 1 = female) Age Education (high school or less, some college, college, advanced degree) Income (16 category ordered measure) Party identification (Strong Democrat to Strong Republican) 660 PS October 2007