Vote Centers as a Strategy to Control Election Administration Costs

The rising costs of election administration in an era of fiscal stress have motivated some local officials to test the feasibility of ideas for reducing election costs while enhancing voter convenience and perhaps even increasing voter turnout. One such pilot project in a suburban community in the South involved replacing precinct-based voting on election day with a vote center that all voters could use regardless of their precinct of residence. A comparison of election costs across two municipal elections showed that replacing precinct-based voting with an election day vote center resulted in substantial cost savings. While there was no statistical difference in voter turnout in municipal elections held before and after implementation of the pilot project, voters were highly satisfied with the convenience of the vote center as well as other aspects of their voting experience. The findings suggest that an election day vote center can be a viable strategy to control election costs and enhance voters’ perceptions of the convenience of voting.

Election administrators operate in a challenging environment. As gatekeepers of the electoral process, they face high expectations for preserving and enhancing the integrity and accuracy of that process. Fulfilling those expectations in a climate of distrust, political polarization, and perceived partisanship requires a special kind of professional dedication and commitment. The job of counting every vote and maintaining confidence in a system in which every citizen's vote counts is a daunting one in the best of times, but in an era of financial stress for many local governments, election administrators need to be more creative and innovative than ever.
One of the principal concerns among election administrators is how to pay for the rapidly rising costs of elections due mainly to the unprecedented level and extent of federal mandates and to a lesser extent, the growing number of state policy changes. Figuring out how to pay for implementation of these requirements as federal funding declines or ends and as local fiscal pressures become more acute has created what Montjoy (2010) describes as a "perfect storm of election finance" (p. 868). Indeed, among all of the challenges that confront election administrators, there is general agreement that election finance is the "critical component underlying most issues" (Fischer, 2008;Montjoy, 2010, p. 867).
The 2002 Help America Vote Act, for example, provided funds for several types of election-related needs that included statewide consolidation of voter files as well as about two thirds of the costs for new voting machines, optical scanners, and laptops for electronic voter registration data files. That equipment is now aging and it is unlikely that additional federal funds will be available to replace these machines (Stein & Vonnahme, 2009). With local budgets squeezed by the "new normal" that prevails in the wake of the Great Recession, the issue of how to pay for an operable inventory of voting equipment to conduct elections as federal funds disappear is just one of many financial headaches that have compelled some election administrators to evaluate strategies to control election costs.
While the 2000 presidential election spawned broader scholarly interest in many of the issues related to election administration, very few studies focused on the actual administrative costs of elections (Montjoy, 2008(Montjoy, , 2010. Research on the cost of election administration is scant (Stein & Vonnahme, 2011). Most often, election "costs" have been examined in the context of what voters experienced after changes in election rules, policies, or voting technologies. Only a few studies have examined the financial costs of election administration (Burden, Canon, Mayer, & Moynihan, 2009;Indiana Fiscal Policy Institute, 2010;Lindback, 2010;Montjoy, 2010;Stein & Vonnahme, 2009). The paucity of research on this aspect of election administration led Montjoy (2008) to assert that "there is an urgent need for additional research to complement existing studies" and especially for "case studies to capture the complexity of impacts from multiple policy changes on the implementing agents" (p. 796).
The purpose of this article is to examine the implications of a vote center pilot project with respect to the costs of election administration, voter satisfaction, and voter participation.
Based on a stakeholder participation process, this pilot project involved securing state legislative approval to replace election day precinct-based voting with an election day vote center, in effect extending the use of an early voting center location on election day for use by registered voters regardless of their precinct of residence. In the case study community, the costs of administration are compared for municipal elections held before and after the policy change. Since any cost savings from vote centers have little value if the price is depressed voter participation, turnout data are examined for the municipal elections held before and after implementation of the pilot project. Likewise, vote centers are not politically feasible if voters are dissatisfied with their vote center experience. An exit poll of a sample of voters measured opinions about various aspects of their voting experience at the vote center.
This pilot project in one community was motivated by the county election administrator's desire to determine whether it would be financially and politically feasible to eliminate, at some point, all precinct-based election day voting in the county and replace it with election day voting at the same 10 vote centers used during the county's early voting period. The thinking was that a pilot test of this idea would permit election commissioners to evaluate the related political and financial consequences and enable them to make a more informed decision about whether they should seek state legislation to enable countywide adoption of the policy. In discussions with state lawmakers about draft legislation to test the idea of replacing election day voting at precincts with vote centers on election day, there was deep concern about how any change in the status quo might deter, repress, or otherwise lower voter participation on election day. A political compromise was reached that enabled the election day vote center idea to be tested in just one city in the county. While a more sophisticated design involving random selection of precincts and random assignment of these to a treatment and control group would yield the strongest type of evidence for any policy impact, the preferences of social scientists do not often prevail among elected officials in debates about changing election procedures.
This article consists of several sections. A brief review of the relevant literature is followed by a description of the process involved in the design of the vote center pilot project. Subsequent sections analyze election administration costs, voter opinions about their vote center experience, and voter participation in municipal elections before and after the pilot project. The study concludes with a summary of the implications of an election day voting regime that allows voting only at the same vote centers used during the early voting period.

Alternatives to Election Day Precinct Voting
Overcrowded, inconveniently located, and poorly designed polling places are a growing problem in many communities. For some jurisdictions, the sheer number of precincts alone present tremendous administrative challenges that entail assuring an accurate vote count, maintaining the integrity of the vote, and paying for an increasingly expensive, laborintensive system of precinct-based voting on election day (Alvarez & Hall, 2008). For example, in Knox County, Tennessee (the location of the Town of Farragut which is the case study community in this article), there are 92 voting precincts that served a 2009 registered voter population of 264,903. While the countywide voter assignment to precincts averaged about 2,700 persons, many local polling places attracted fewer than 100 voters in the November 2008 presidential election. In that election, a total of 915 workers were needed to staff polling places in the 92 precincts on election day compared with the 100 personnel employed to operate the 10 early voting locations during the state's 15-day early voting period. The personnel costs for staffing 92 precincts on election day was just part of the challenge. These individuals also needed to be trained to deal correctly with an increasingly complex set of rules that concern provisional voters, address changes, failsafe procedures, and other election rules (G. MacKay, personal communication, January 29, 2009).
In addition to concerns about the costs of election administration, there are issues related to the safety and security of students and staff in those precincts that use schools as polling locations on election day (Montjoy, 2008). As one county election administrator opined, Is it wise to invite a large number of strangers to wander through a school, individuals who on any other day would be required to check in at a central office and perhaps even be issued a visitor's photo identification? (G. MacKay, personal communication, January 29, 2009) Many election administration costs are fixed and relate to voting equipment, leases for voting stations, printing, postage, voter registration, and maintenance of voter registration files. The fiduciary challenge for election administrators is to identify strategies that might reduce the costs that are controllable while still providing voters with a convenient, safe, and accessible voting experience that does not deter participation. Obviously, local election officials would not condone any cost saving strategies that cut compliance costs or reduce the number of elections.
A potential target for improving efficiency is the traditional election day precinct voting system that may have scores of lightly utilized neighborhood voting stations. Similar to early vote centers, election day vote centers would permit voters to cast their ballots at any number of voting places in a county rather than just the one polling place in the vicinity of the voter's place of residence. The idea is that these vote centers are more convenient and central to voters' work, shopping, school, or travel routes.
The extant research on the principal alternatives to election day precinct voting such as early voting, mail-in voting, and vote centers has focused mainly on how these methods and their various permutations affect voter turnout and perceptions of voter convenience. Very little is known about how the cost of these alternatives compares with traditional election day precinct voting. Most studies find that mail-in voting is popular but research about its effects range from a national study that found no impact on turnout (Fitzgerald, 2005) to a single state study of Oregon that concluded mail-in voting resulted in a 10% increase in turnout (Richey, 2008). The consensus in the literature suggests that early voting just makes the task of voting more convenient for those who would have voted anyway and has little or no impact on turnout (Gronke, Galanes-Rosenbaum, & Miller, 2007;Richey, 2008;Stein & Gracía-Monet, 1997). An exception is Burden et al. (2009) who found early voting actually decreased turnout. When the administrative costs are not mentioned at all, the literature on mail-in and early voting imply that they are likely to vary; however, actual cost analyses of these alternative voting methods are rare (Burden et al., 2009). Gronke, Galanes-Rosenbaum, Miller, and Toffey (2008) described this as a "dearth of research on campaign effects [and] election costs" (p. 437).
Majorities of voters in Tennessee voted early in the last two presidential elections (57.9% in 2008 and 58.7% in 2012). The question for local election commissioners was whether it might be possible to realize savings in administrative costs without increasing voter inconvenience or deterring participation if the same vote centers used during early voting were employed on election day? The potential gain in financial efficiency is the central appeal of Election Day vote centers.
An election day vote center is a polling place where any voter registered in the jurisdiction may go to vote. Unlike election day precinct voting, there is no wrong place to vote. Typically, these vote centers utilize electronic voting equipment with appropriate ballots; electronic poll books on PCs provide access to registration data (Doyle, 2008). Montjoy (2008) pointed out that vote centers may help reduce voter confusion about where to vote and may help reduce election administration costs compared with precinct-based voting.
The idea is that having a smaller number of centrally located, accessible vote centers may enhance the convenience of voting and decrease expenses associated with staffing, equipping, and operating numerous precinct polling places (Stein & Vonnahme, 2008). Larimer County, Colorado, was among the first jurisdictions to implement this concept in 2003. It replaced 143 separate polling places with 22 vote centers that utilized technology that allowed election officials to access voter records to assure that individuals voted only once (Rokita, 2005). Juenke and Shepherd (2007) examined the turnout effects of vote centers in Colorado but did not find any significant impact on turnout. Stein and Vonnahme (2008) analyzed individual voter histories and found that election day vote centers increased voter turnout generally but that the largest effect was among infrequent voters. Scheele, Losco, Crawley, and Vasicko (2008) compared turnout in Indiana municipal elections that used traditional precinct-based voting and those that utilized vote centers. After accounting for differences in the competitiveness of elections, they concluded that vote centers had a minimal impact on turnout but that large majorities of voters, especially early voters, thought that vote centers were as convenient or more convenient than their old precinct polling place.
As noted, much less is known about whether election day vote centers result in more efficient election administration than precinct-based voting. Stein and Vonnahme (2009) compared election administration costs in Colorado counties that utilized precinct-based voting and Election Day vote centers during the November 2008 election. They found that election costs in counties with vote centers averaged US$2.83 less per registered voter than counties that utilized precinct-based voting on election day (US$7.92 vs. US$10.75 per registered voter). They concluded that election day vote centers can lead to slightly lower election administration costs and to a higher level of voter satisfaction and turnout.
Based on an analysis of the performance of vote centers in three Indiana counties, the Indiana Fiscal Policy Institute (2010) compared the estimated costs associated of conducting an election with a system of vote centers in each of the state's counties with the costs of a precinct-based system on election day. The study concluded that vote centers would produce significant savings for all counties but particularly for those jurisdictions with smaller numbers of registered voters per precinct (Indiana Fiscal Policy Institute, 2010). The projected savings ranged from 20% to 56% across counties due mainly to reduced expenditures for poll workers' compensation and voting machines as fewer personnel and less equipment would be needed to conduct an election (Indiana Fiscal Policy Institute, 2010).
Montjoy (2010) examined election costs in Weld County, Colorado, which implemented vote centers between the 2004 and 2008 elections. Although the county reduced the number of polling places by two thirds, it did not realize any reduction in costs as measured by the costs incurred per vote due to several jurisdiction-specific circumstances. These included the decision to increase the pay of poll workers and the costs associated with processing the large increase in the number of absentee votes cast in 2008.

The Farragut Pilot Project
The increasing frequency of voters who appear at the wrong precinct to cast their vote on election day, the growing proportion of voters who vote early at convenience voting centers, and uncertainty about local finances prompted local election officials in Knox County, Tennessee, to examine alternatives to election day voting by precinct. According to the county's Administrator of Elections, evaluating the merits of election day vote centers was motivated by the findings from the Larimer County, Colorado, vote center experience, the desire to make voting as convenient as possible, and the need to utilize resources efficiently (G. MacKay, personal communication, January 29, 2009). Local election officials wanted to know what voters thought about the idea of using vote centers on election day, whether it would affect voter participation, and what if any savings might be realized in election administration costs.
As discussions about the idea evolved, local officials realized that work would be required on two fronts simultaneously: the state legislature and county voters. State law would need to be changed to allow the county to implement an election day vote center pilot project as existing statutes required the early voting period (when voters cast ballots at vote centers or at the election commission offices) to begin 20 days before and to end 5 days ahead of election day at which time voters were required to cast ballots in the precinct of their residence. Second, local officials wanted to determine what county voters and local stakeholder groups thought about the idea of replacing precinct voting with election day vote centers.
On the legislative front, work with the county's state representatives resulted in the introduction of a bill that was passed eventually by the state's General Assembly to provide for an election day vote center pilot project to be implemented in 2009 for a municipal election held in any city that agreed to participate (Convenience Voting, 2008). The legislation specified that the period for early voting in a pilot project would begin on the 20th day before election day and would continue up to election day, excluding Sundays and state holidays. The law also stipulated that a super-majority of at least four out of five county election commissioners would be required to approve a city's participation in the pilot program to assure bipartisan support (Convenience Voting, 2008).
This enabling legislation differed in two major ways from the existing election regime. First, instead of ending no less than 5 days before election day as required by state law, early voting would be extended through election day. This amendment to the original bill added three additional days of early voting (excluding Sundays). Second, the new statute granted some discretion to the Knox County Election Commission to establish the location and number of convenience voting centers to replace election day precinct polling as long as there would be at least one vote center for every 25,000 registered voters. In effect, this meant that the polling places used for early voting could be the locations for election day vote centers.
Several methods were used to help determine what county voters thought about the idea of replacing precinct voting at vote centers on election day. These included a preelection focus group, a pre-election telephone survey of a sample of county voters, and an exit survey of voters in the 2008 election. The focus group was comprised of community leaders, election workers, and various interest group stakeholders (Gibson et al., 2008). The consensus among participants was that the idea of replacing precinct voting with vote centers on election day merited at least a trial. The focus group feedback was instrumental in refining questions for a subsequent telephone interview with a randomly selected sample of the county's registered voters. Questions in this survey measured what registered voters thought about the idea of casting ballots at vote centers rather than at precinct polling places on election day and what features and location criteria of a polling place were important. Telephone interviews were conducted between November 11 and 30, 2007, with a random sample of 400 registered voters in Knox County. The error margin in the survey was ± 5.0%.
When asked their opinion about the idea of replacing the election day polling places in each precinct with a system of centrally located voting centers that would operate like those used during early voting where registered voters could vote at any county voting center, the majority of respondents (53.7%) indicated that they favored this idea, 19.9% opposed it, 22.7% neither favored nor opposed the idea, and 3.7% of the 397 respondents were not sure. Neither political ideology, party identification, nor any other demographic feature of respondents was associated with or helpful in explaining the distribution of opinion on this question. Among the 79 respondents who opposed the idea, the reason offered by the largest proportion (27%) was that they thought vote centers would "make it harder for some voters to participate." Another 20% of those opposed to vote centers said they held that opinion because they "do not like change," and 18% thought the idea would "create confusion among voters." The remaining respondents provided reasons for opposition that were too diverse to code into distinct categories.
Among the 213 respondents in the telephone survey who favored the vote center idea and who provided a reason for their opinion, about half (49%) said they thought voting "would be more convenient." Another 18% favored the idea because they thought it "might increase participation" and 14% thought it would "reduce confusion about where to vote." Some respondents (7%) favored the idea because they thought it might "reduce the costs of elections." These findings encouraged local officials to conduct a more extensive exit poll of voters in the 2008 presidential election to confirm what these voters thought about the election day vote center idea.
An exit survey was designed and administered to a randomly selected sample of 1,574 Knox County voters in the November 2008 election (± 2.6% error margin). This survey was administered by trained poll workers to a random sample of voters at each of the vote centers during early voting and at all of the precinct polling places on election day. The profile of respondents mirrored the proportion of voters who voted early and on election day (57% and 43%). The wording of the relevant question in the exit poll differed slightly from the question asked in the 2007 telephone poll to reflect the statutory language that had been passed by the state legislature. Voters were asked whether they would be in favor of extending early voting at vote centers through election day and doing away with precinct-based voting on election day. Of the 1,560 voters who answered this question, 920 (59%) answered "yes," 301 (19.3%) answered "no," and 339 (21.7%) were "not sure." No statistically significant differences in this opinion occurred with respect to the gender or educational attainment of survey respondents. However, African American voters were statistically much more likely (71.8%) than White voters (56.0%) to favor extending early voting through election day and replacing precinct voting with vote centers (Cramer's V = .093, p = .00). On the contrary, older voters were statistically somewhat less likely to favor extending early voting through election day and replacing precinct voting with vote centers (Cramer's V = .122, p = .00). A majority of voters in each age group favored the proposal, but the proportion of support declined from a high of 63.7% in the 26 to 35 age group to 53.1% among those above 65 years of age.
Another statistically significant difference in support for extending early voting through election day and replacing precinct voting with vote centers occurred among those who voted early and those who voted on election day (Cramer's V = .30, p = .00). Among the survey respondents who voted early (about 60%), 70.3% favored the idea, 9.8% did not favor it, and 19.9% were not sure. Among the voters who cast their ballot on election day at a precinct polling place, 44.0% indicated that they would favor extending early voting through election day and doing away with precinct voting, 31.8% did not favor the idea, and 24.1% were not sure. Analysis of the responses indicated that while comparatively more voters above 65 years of age tended not to support the change, more voters in this age group actually voted early at a vote center (72.4%) than any other age group.
County election officials considered the findings from these different methods as sufficient evidence of voter support to move forward with implementation of a pilot project in 2009. While both cities in Knox county had local elections scheduled for that year, the Board of Mayor and Aldermen in Farragut was more receptive to the idea than the Knoxville City Council. After approval by the Town of Farragut, the county election commission chose this community to be the site of the pilot project for the April 2009 municipal election. This city is a suburban bedroom community that had a population of 20,689 people and 15,963 registered voters in 2009. The median household income was US$95,106 and more than half of those 25 and older (56.5%) had a bachelor's degree or higher. Minorities comprised about 8% of the population (U.S. Census, 2010).
The Farragut town hall, the designated site for early voting in all other elections, would continue to be the early voting location for the pilot project. The principal change in the pilot project concerned where voters would cast their ballots on election day. In previous elections, after the end of the early voting period, the entire voting infrastructure at town hall was dismantled, packed, and moved to a high school less than a mile distant where it was re-assembled, tested, and following state law, configured to create two precinct polling stations to accommodate voters who lived in each of the town's two precincts. Election day voting in the April 2009 election would occur instead at the Farragut town hall, the same location as early voting.
The traditional election day precinct-based voting regime in Farragut involved drayage costs, additional voting equipment, and more election workers to staff two separate polling stations at the local high school. In contrast, the pilot project would utilize the same basic equipment and staffing configuration on election day as during early voting with ballots programmed to record the resident's precinct of residence using the county's e-Slate electronic voting machines.
New costs incurred by the pilot project involved advertising both the extended early voting opportunity and the new town hall location for election day voting. This effort involved advertisements in the local media and postcard mailings to each registered voter that described the times of operation for early voting days at town hall and the use of that facility for election day voting.

Election Administration Costs
To help capture all election expenses, cost data were collected and assigned to one of the three phases of election administration: setup, operation, and takedown. Actual costs were itemized for each of these election administration phases. To determine whether the pilot project resulted in any savings, election administration cost data for 2009 were compared with the costs incurred in the previous 2005 Farragut municipal election. Scheele et al. (2008) found that the level of competitiveness of a local election affects turnout and associated administrative costs. An analysis of the competitive nature of the races in the two Farragut elections concluded that both involved a highly competitive mayoral race between an incumbent and a challenger. In the 2005 election, the incumbent won, but in 2009, voters elected the challenger to be the next mayor. Both elections involved the selection of an aldermanic candidate to represent each of the town's two precincts (66N & 66S). In the 2005 election, both aldermanic contests were contested, however in 2009, only the 66N precinct race was competitive. The candidate for the 66S precinct seat was unopposed. On balance, the level of electoral competitiveness in the two mayoral campaigns was generally comparable. Table 1  In both elections, the cost for election workers' pay was the single largest expense. In 2005, poll worker's pay accounted for 81% of all election administration costs and overtime pay alone comprised 26.5% of all salary costs. In 2009, poll worker's pay accounted for 72% of all expenditures and overtime pay comprised 19.9% of all salary costs. A total of 26 election workers were trained and hired for the 2005 election in Farragut but not all of these personnel were employed full-time for the entire 15-day early voting period. About a dozen part-time workers were needed in that election to help transport voting equipment to the high school after early voting and to operate the two voting precinct stations for the 12-hr election day voting schedule.
These additional staffing expenses were unnecessary at the town hall vote center in 2009 because fewer but more experienced full-time staff were able to manage operations at the vote center (G. Mackay, personal communication, June 12, 2009). Even with three additional days of early voting, fewer election workers translated into lower overall expenses for salaries and training (G. Mackay, personal communication, June 12, 2009). The higher set-up costs in the 2009 election are attributed to the expenses associated with postage, printing, and advertising the change in the voting period and election day voting location.
Cost per vote cast is a useful metric for comparing administrative costs across elections. Table 2   These comparisons show that the 2009 pilot election day vote center in Farragut had much lower level of staffing costs compared with the 2005 election even though the 2009 election included three extra days of early voting and 71 fewer voters. That a significant difference was observed in the lower overall cost of election administration suggests that an election day vote center can be an important strategy to help reduce the costs of election administration. This finding may have particular import for the growing number of communities that confront the unbudgeted costs connected with local recall elections (Greenblatt, 2009).
Encouraging voter turnout is an important objective for election administrators. More voters can result in a lower cost per vote particularly when no changes are required to equip and staff an election day vote center compared with that of early voting. Subsequent sections examine voter participation in the 2005 and 2009 elections and what voters thought about their experience at the town hall vote center.

Voter Turnout
Conceptually, a vote center might be expected to enhance the level of voter participation by making the act of voting less "costly" by reducing the time and inconvenience associated with this form of civic engagement. With the exception of the Stein and Vonnahme (2008), previous studies found little evidence to indicate that vote centers encouraged higher levels of voter participation. In local elections, evidence of any effect is particularly difficult to detect considering the large number of variables that affect voter turnout and the overall trend of declining participation in these elections. The competitiveness of particular races, the salience of various issues, and changing voter demographics are just a few of these factors. On the contrary, a vote center may be similar to many other electoral reforms that simply helped to make the act of voting more convenient for those who would have voted anyway. Table 3 shows the number of voters in the 2005 and 2009 Farragut, Tennessee, municipal elections. Although 1,336 more individuals were registered to vote in 2009 than in 2005, there was no statistically significant difference in voter turnout. The large majority of voters (77.28%) who cast their ballots before election day in 2009 indicated that early voting continues to grow in popularity among those who vote in a municipal election. By contrast, about 60% voted early in 2005. In the 2009 election, more than one fourth of all voters (28.0%) took advantage of the pilot project's extended early voting period. That level of participation suggested that the election commission's publicity effort succeeded in reaching the attentive voters in Farragut with news of the expanded opportunity for early voting. Table 4 shows the turnout rates in each of the town's precincts. While overall turnout was 1.81 percentage points lower in 2009 than in 2005, all of this difference was attributable to the lower level of electoral participation by residents of the 66S precinct which did not have a competitive aldermanic election in 2009. As participation among voters in the 66N precinct actually increased slightly (by +.63% percentage points), it is reasonable to conclude that the lower turnout in the 66S precinct (−4.70 points) was due mainly to the non-competitive nature of the aldermanic election (only one non-incumbent candidate was on the ballot) and not to anything associated with voting mechanics of the pilot project. These turnout data suggest that neither an extended early voting period nor the opportunity to cast a ballot on election day at a vote center had any effect on voter participation in 2009. This no-impact finding suggests that adoption of an election day vote center strategy countywide would not have any significant effect on voter turnout.

Voter Experiences at the Vote Center
To determine what voters thought about their experiences at the vote center, an exit questionnaire was designed and poll workers were trained to distribute the instrument to a randomly selected sample of voters. Following a protocol in which every third voter was asked to complete the questionnaire after voting, a total of 670 voters agreed to complete the survey questions. Responses were collected from voters who cast ballots at the Farragut town hall from March 18th through April 6th (early voting) and on April 7th 2009 (election day). This sample represented about 29% of all voters with a sampling error of ± 2.7% at the 95% confidence level. Excluded from the survey were the 19 individuals who voted at the county courthouse and the 57 persons who cast absentee ballots. The sample generally represents the proportions of individuals who voted during regular early voting, extended early voting, and on election day. Table 5 shows that an overwhelming majority of voters (96.0%) rated the convenience of the Farragut town hall as a vote center as "excellent" while 4.0% rated it as "good." Not a single respondent rated it as fair, poor or was not sure about its convenience. Among those who voted on election day, the group for which the vote center most likely represented a new polling place, 98.3% rated the convenience of the vote center as "excellent," an even higher proportion than that recorded for those who voted early. While no opinion data from voters in 2005 are available, if it is assumed that voters in that election were equally satisfied with their election day voting experience at the high school, the lower administrative costs of the pilot election day vote center still make this an attractive strategy for local election officials. There were no statistically significant opinion differences among voters concerning the convenience of the vote center with respect to their gender, race, age, or education characteristics. Table 6 shows that voters uniformly rated the convenience of the vote center as excellent regardless of differences in their attributes.
One reason voters may have rated the convenience of the Farragut vote center so highly was that their wait time to vote after arrival at the polling place was typically less than 1 min, regardless of when they voted. As Table 7 indicates, almost 80% of voters waited less than a minute to vote after arrival at the polling place. For the remaining 20%, the wait to vote was between 1 and 5 min. For those who voted on election day, three fourths of voters waited less than a minute while one fourth waited between 1 and 5 min. A municipal election with light voter turnout is not an ideal test of the performance of a vote center in terms of wait times to vote, but the respondents' very favorable rating of the convenience of a vote center augers well for its use in lieu of two separate precinct polling stations on election day.

Implications of the Findings
This case study represents a modest but important contribution to the literature on vote centers in that it finds evidence that an election day vote center has potential to reduce the costs of election administration compared with a precinct-based system. In inflation-adjusted dollars, those savings amounted to a 14.5% reduction, despite the fact that the pilot project included three extra days of early voting. The principal source of savings consisted of reduced costs for poll workers' compensation, particularly overtime pay. The vote center required fewer poll workers and reduced the recruitment and training workloads associated with staffing two separate polling stations on election day.  The cost reduction realized in the vote center project represents a conservative estimate of economic benefit. Personnel costs typically account for about 87% of total election costs in Knox county. In the 2008 presidential primary election, for example, total costs for just election day voting amounted to US$277,986 significantly higher than the US$243,377 total costs for the early voting period. (Gibson et al., 2008). Assuming that the county did not have to hire and train some 800 poll workers for election day voting would reduce personnel costs by about US$192,000 assuming that staffing in the 10 vote centers would double on election day. This translates into a total countywide election cost reduction of about 36%. It seems reasonable to suggest that other small communities in states with comparable early voting periods could expect to achieve similar or larger savings in election administration costs depending on the importance of various factors that drive their election costs.
The findings from this case study indicate that local election officials should not expect election day vote centers to result in a substantially different level of voter participation than in previous, comparable elections. Voter turnout was essentially unchanged in this study. While some might argue that any departure from the historical pattern of decline is a positive indicator, there was no evidence that having an election day vote center figured into voters' decision to go to the polls or not.
Regardless of their socioeconomic status or race, voters rated their experience at the vote center in very positive terms. The extent to which voters in other communities have a similar experience will depend on the extent to which local election officials establish vote centers in accessible locations with adequate parking-facilities that figure prominently in the voters' calculus of convenience.
In conclusion, election day vote centers merit scrutiny by local election officials as a strategy to control the costs of election administration. Even though overall voter turnout may not change, citizens satisfied with the convenience of their vote center experience at least have no facility-related reason not to return to participate in the next election. The finding that elections can be administered for a significantly lower cost without deterring participation should elevate the salience of this strategy on the policy agendas of both state and local governments in an era of fiscal austerity.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research and/or authorship of this article.