Los Angeles Polling Place Accessibility Study

Primary Election
March 2, 2004

Matt A. Barreto
University of California, Irvine
Center for the Study of Los Angeles

Mara Marks
Loyola Marymount University
Center for the Study of Los Angeles

Nathan D. Woods
Welch Consulting
Center for the Study of Los Angeles

Precinct Quality and Voter Turnout: Race, Income and Civic Participation

Prepared for delivery at the Midwest Political Science Association Conference
Chicago, IL April 15, 2004

* The authors would like to thank Fernando Guerra, Director of the Center for the Study of Los Angeles, and Xochitl Marquez and Haven Perez, researchers at the Center, who's support was integral in coordinating and completing this research project. We would also like to acknowledge the enthusiastic participation of the 90 undergraduate students who carried out the research at 960 precincts throughout Los Angeles.

Introduction
A citizen's decision to go to the polls on Election Day and cast a ballot is a longstanding preoccupation of political practitioners and political scientists alike. Even in this highly contested upcoming presidential, a large number of eligible citizens will not participate in the democratic process. Since the 1960s, scholars have documented the growing number of non-voters, and wondered why turnout has been on the decline (Teixeira 1987; Wattenberg 2002). While previous studies have cited declining trust in government, uncompetitive races, too frequent elections, changing demographics, and depleting social capital, few have focused explicitly on the costs associated with voting (Rosenstone and Hansen 1993). Many of these studies focus on the costs of voting and examine how changes in registration laws or early voting procedures might reduce the costs of voting (Brians and Grofman 2001; Highton 1997). Only one previous study has examined the voter's physical relationship to the voting precinct as a possible determinant of turnout (Gimpel and Schuknecht 2003).

While the aim of this paper is not to explain declining turnout over time, we are interested in the costs of voting that citizens may incur via their polling place. We argue that not all precincts are created equal: those that are less accessible, of lower quality or have less informed poll workers, will depress voter turnout. Further, these low-quality precincts may not be randomly distributed within a city or state, and rather may be more prevalent in low-income and minority neighborhoods. This paper investigates the extent to which precincts in Los Angeles vary, the geographic distribution of low-quality precincts, and the relationship between precinct quality and voter turnout.
In assessing the relationship between precinct quality and voter turnout, this paper proceeds in four sections. First, we review the relevant literature on the costs of voting. Second, because of the uniqueness of our data, we review the design and implementation of the polling place accessibility research project. Next, we detail the findings concerning the quality and accessibility of precincts in the study area, Los Angeles, California. Finally, we test the relationship between precinct quality, race and income of the precinct neighborhood, and precinct level voter turnout.

I. The Costs of Voting
Although few citizens consciously conduct a cost-benefit analysis before deciding whether or not to vote, all potential voters decide whether it is worth their time to travel to the polling place, decide how to vote, and cast a ballot. While many factors influence a citizen's decision to vote on Election Day, they must first overcome some hurdles or costs associated with voting. Registering to vote, becoming informed about the issues and candidates, learning where the voting precinct is, and finally taking the time out of the day to vote. Generally, these costs can be divided into two categories, institutional and individual. Institutional costs are those associated with the rules and regulations of voting such as registration requirements while individual costs are those that impact individuals differently such as gathering information.

The classic work on voting costs, Anthony Downs' An Economic Theory of Democracy (1957) is baseline for research on turnout. Downs basic argument is that the benefits of voting are so low that even relatively small costs will preclude rational citizens from casting a ballot: "when voting is costly, its costs may outweigh its returns, so abstention can be rational even for citizens with party preferences. In fact, the returns from voting are usually so low that even small costs may cause many voters to abstain," (1957: 274). Efforts to determine the potential benefits of voting are also costly, requiring the citizen to determine which party or candidate would increase their expected utility and then to calculate the likelihood that theirs would be the deciding ballot. If the citizen concludes that there are benefits to voting, they then confront the costs of registering to vote, traveling to the polling place, and casting the ballot. Faced with these costs and recognizing that they are unlikely to cast the deciding ballot, most citizens ultimately decide that voting is not worth their time. However, many citizens do vote even though they aren't casting the tie-breaking ballot. This may be the result of what Downs, and later Riker and Ordeshook (1968) describe as a sense of civic duty or democratic pride. While the benefits to voting are low, so too are the costs and even when the costs outweigh the benefits, many citizens participate in an election because they feel it is a civic duty. However, if the costs escalate or the benefits diminish, even these good citizens may decide it is not worth their time to vote.

More recently, Blais has reexamined the costs of voting and concluded that in modern democracies, the costs of voting are extremely low and unlikely prohibitive (2000). In listing the costs of voting, Blais notes that the process of voting is relatively straightforward: "citizens have only to answer a short questionnaire to become registered on the electoral list, to go to a polling station that is usually located close to where they live; and to indicate on a ballot which party(ies) and/or candidate(s) they wish to support. These activities are supposed to require very little time and effort," (2000: 84). However, the polling station may not be so convenient as Blais supposes. Research by Gimpel and Schuknecht demonstrates that polling place location can negatively impact turnout (2003). Their study considers the placement and geography of polling places relative to the population they serve and finds that as suburban voters live further away from their precinct, they are less likely to vote. The authors note, "some precinct locations are more accessible than others, and for the less accessible ones, at least some people will feel that the costs to get there outweigh any benefit," (472). Although their research only examined one facet of precinct accessibility -- distance from home -- Gimpel and Schuknecht bring renewed attention to the costs of voting. They argue that because accessibility "has rarely been one of the convenience factors subject to measurement," future research on voter turnout "could benefit by using innovative methodologies to examine geographic accessibility of precinct polling places," (472-3). While Blais contends that the costs of voting are low, his own research confirms that increasing the costs for voters at the precinct can reduce turnout. He estimated that increasing the time it takes to vote from 15-30 minutes to 45 minutes would result in lower voter turnout (2000).

Along these lines, scholars concerned with the calculus of voting have examined potential demobilizing factors that come up on Election Day. Knack investigated the impact that bad weather, namely rain, had on the turnout rates of Democrats and Republicans (1994). He found that while some voters are committed to voting and little will stop them, for other potential voters without a strong sense of civic duty, rain steadily decreases voter turnout. Merrifield also found similar results for inclement weather and turnout (1993). Finally, Katosh and Traugott (1982) reported that access to the polling place, measured by number of hours open is correlated to voter turnout with lower voting at polls that are less accessible.
Prior to 1965 polling places were known to be cost-incurring and the Voting Rights Act (VRA) purposely eliminated Election Day costs such as poll taxes, literacy tests while also improving access to precincts in low income and minority communities. Since then, little attention has been paid to the polling place, however the events in Florida during the 2000 Presidential election brought to the forefront discrepancies in voting rights. Some voters in Florida protested that they were dropped from the voting rolls, their precinct location had been changed without notice, or that they were asked to present identification before voting (Navarro and Sengupta 2000).

Increasingly, election officials concerned about decreasing levels of turnout have recognized obstacles to voting and attempted to make the ballot more accessible. Most notably, Oregon has gone to an exclusive system of vote by mail while Arizona, Michigan and Iowa have tried Internet voting experiments to give voters more flexibility and reduce the costs associated with polling places.

Here we examine the specific costs associated with finding the polling place, navigating the polling place, and potential barriers to voting at the polling place. For example, if the voting precinct location changes from election to election, voters may be unfamiliar with where they should go to vote. Similarly, if a voter pressed for time cannot find a parking spot near the precinct, he or she may drive around the block twice, and then decide to go home. At the aggregate level, such precincts may experience lower voter turnout. While these are just two simple examples, our study brings to light many new institutional costs of voting impacting access to the ballot and effecting turnout rates.

II. Overview of the Project
In March 2003 the Center for the Study of Los Angeles at Loyola Marymount University launched a pilot study of polling place accessibility in precincts randomly selected from throughout the city. Two additional waves of the study were completed, in October 2003, and March 2004 . This research focuses on how the physical characteristics of polling places affect their accessibility as well as the ease and convenience associated with the voting experience in Los Angeles. The purpose of the polling place accessibility study was two fold: first, to get an overview of precinct quality and accessibility for the city of Los Angeles as a whole, and second to determine if differences in quality were related to race and income characteristics of the surrounding neighborhood. For example, several polling places in Los Angeles were hard to find, did not have addresses clearly listed, and offered limited parking. Polling places varied widely in their size, comfort of the waiting area, the number of machines available, and the knowledge of poll workers. Further, there were some notable differences within the city based on the demographic profile of the precinct neighborhood. . The results reported below show the raw results of the March 2004 citywide study and summarize some of the key variables for which data were gathered. Taken together, these results reveal important insight into the voting experience in Los Angeles and how widely it varies among polling places in the city.

We attempted to examine three general criteria of precinct accessibility and quality: (1) are polling places easy to find; (2) are polling places easy to use and comfortable; and (3) are there any barriers to voting in polling places? However, before we discuss the results, we offer a review of the methodology and survey instrument that was used in carrying out this research project.

How Was the Survey Administered?
In March 2004, the third round of the study, we recruited approximately 90 undergraduate students from Loyola Marymount University to carry out the surveys throughout the city of Los Angeles and five neighboring comparison sites. Researchers at the Center for the Study of Los Angeles created a "check-list" of criteria to gauge how accessible and voter-friendly a polling place is. This checklist reflects previous scholarship, our personal experiences over the years at various precincts, students' observations in previous rounds of the study, as well as discussions with precinct workers, and high-ranking officials in the Los Angeles City Clerk's office and the Los Angeles County Registrar-Recorder's office. The checklist includes straightforward questions about the polling places designed to be easy for the students to administer . Questions were concise, and the possible answers were almost always Yes/No or High/Medium/Low depending on the category of the question, leaving little room for error on the part of the survey administrator. As a further safeguard to ensure consistency in completing the checklist, two mandatory training sessions were conducted one week prior to Election Day and on the morning the polls opened.

Generally, we sought physical descriptions of the polling places, both inside and outside, and how the visibility of the polling places to passers by. We asked whether flags or banners helped mark the location as a voting precinct, whether adequate outside lighting was present to illuminate the polling place at night, and whether the polling place had been previously used in an election or if it was a new location for voters. In addition, we wanted to determine how knowledgeable poll workers were with respect to voting rules and regulations such as the need for identification and the right to vote a provisional ballot (for full survey see Appendix A).

Prior to the March 2, 2004 election, we acquired a complete list of polling places and addresses to be used by the County of Los Angeles, and divided them into the 42 zip code areas throughout the city. In addition to Los Angeles, we selected five neighboring communities to study as comparison sites. These included Compton, Inglewood, Beverly Hills, Santa Monica/Marina del Rey, and Bell Gardens/Huntington Park. There were roughly 1,700 precincts in use in Los Angeles on Election Day, and student teams were sent to a random selection of 1,350 precincts. Pairs of students were assigned a random list of precincts within a zip code area or comparison site. Students were asked to complete 20 surveys, for a total of 840 precincts citywide, about 50 percent of all precincts in the city of Los Angeles. When the election had ended, the 42 teams of students had completed 960 surveys, about 23 per team, resulting in a +/- 2 percent confidence interval for our findings.

During the student training sessions, we explained the relevance of the project to academics and practioners, and stressed how important it was that each team provide reliable data. We carefully explained to the students each of the 29 questions on the checklist to ensure that all surveys were filled out using the same guidelines. Students were given their address list one week before the election and asked to map out directions to each precinct to increase the chances that they would find each precinct to which they were assigned, maximizing the checklist completion rate. In addition, each team had a camera, and photographed the best and worst practices among polling places in Los Angeles to better document the research. On Election Day, we ensured that all student-teams were equipped with the complete address and name of the polling places to be surveyed and street maps of Los Angeles and surrounding communities.

III. Precinct Quality and Accessibility in Los Angeles
As noted above, our study focused on three areas of precinct quality that voters might encounter: (1) are polling places easy to find; (2) are polling places easy to use and comfortable; and (3) are there any barriers to voting in polling places? While some of the results reported may seem insignificant, there are over 1.5 million registered voters in the City of Los Angeles, creating the opportunity for a considerable number of voters to be disadvantaged even when just 10% of precincts are flawed. In addition to marking "yes" or "no" check list, students were asked to take notes on best and worst practices that they encountered during their research, and we include these observations below.

Are Polling Places Easy to Find?
The first several queries were intended to assess how easy it was to actually find the polling place. The frequencies reported in Table 1 indicate over 20% of polling places did not have street addresses clearly posted outside, making it potentially difficult for voters with limited time to track down the polling place. Only 60% were adjacent to a major street, which could create problems for voters using public transportation. While 69% of polling locations were described as "readily visible" from the street, one in ten were not visible and two in ten were only somewhat visible. While all polling places receive flags and signs to display, about one quarter of precincts did not have a flag or banner displayed in a way to help identify the location as a voting precinct. About one in four polling places did not have adequate outside lighting to illuminate the precinct for those citizens coming to cast a ballot in the evening. Finally, most polling locations were described as easy to find (84%) with only one in six locations described as somewhat or very hard to find. This may have been due to the fact that students were asked to map out directions to the precincts; however, voters also receive precinct addresses before the election in their voter pamphlet and have an opportunity to look up directions. In full, the data reported here, and the student observations below reveal that many differences exist in precinct quality.

[ Table 1 ]
Student observations:

The polling place stood out and there were plenty of signs directing the voters toward the precinct
Apparently relocated to Dunken park. Couldn't get in, had to call desk, no relocation sign.
Had signs in red, white, blue and balloons. Good parking, easy to spot.
The site was hidden behind various barriers and gates and could not be seen from the street
It was a religious school. It had a sign that was very clearly posted that stated, "This property closed to the public" We took a picture of it.
Although signs were posted there were 2 doors, both of which were closed.
"Beware of dog" sign was scary - it was right next to the "Polling Place Here" sign - did not know which to believe
No parking available. We saw 2 people drive away.
Very large parking lot - good!
The parking was all metered
Had to get parking validated.
Filming a movie so lots of traffic, hard to get to
Too many stairs - wheelchair would never be able to get up there.

Are Polling Places Easy to Use and Comfortable?
The second general set of questions attempts to ascertain the ease of use and comfort of the polling places. One-third of polling places did not offer adequate parking, and about 20% were not fully handicap accessible. Both of these findings are particularly troubling given that they may impose barriers to accessing the polling place. Inside, there was wide variance in the size of the polling places and the availability of seating areas. While 19% of precincts were described as "very large," an additional 18% were found to be very small . Related, there were large discrepancies in the waiting area inside precincts as evidenced in Table 2. Half of all polling places had considerable waiting areas with chairs and sofas or large standing areas available for voters, however this stands in contrast to the other half of precincts in Los Angeles. Almost one in six precincts had no extra room, while another one in three had just a small standing area. Taken together, the figures reported below indicate wide variety in the comfort and ease of voting in Los Angeles.
Further, a quarter of polling places had additional amenities present, such as coffee or donuts. Almost two-thirds of the precincts surveyed did not have access to a restroom clearly designated. More troubling, we found 12% of polling places did not have adequate lighting inside to read and mark the ballot. Although Los Angeles County transitioned away from punch card machines in 2004, the ballots were almost identical as before and featured very small print and small circles to be marked with an ink pen. The difficulties in Florida in 2000 highlighted the importance of being able to accurately read and interpret the ballot to ensure that each vote is counted. While the first set of criteria demonstrated that not all precincts are equally visible and accessible across the city, this set of criteria has shown considerable disparities inside the polling places themselves. Inadequate parking, scarce lighting, small rooms and lack of waiting areas characterize about one out of four precincts in Los Angeles, even as others are large, have couches, and serve coffee to waiting voters.

[ Table 2 ]
Student observations:

Homeless people laying next to entrance.
Very posh and swanky; a Bel Air club with ocean view behind every voting booth. Fragrance smelled good.
Had a waiting area. Funny smell.
It's a hotel a small hotel with a small voting room. Smells like older people. No parking.
Funeral was taking place in the church. Polling place was in a building next to the church. Funeral blocked the entrance of the parking lot once the service was over. Police officer was very rude to people in the parking lot.
Terribly inaccessible, got lost, had to walk all over, people complained, barely marked. A women coming out said they were disorganized.
Piano player in the lobby
Big school, but the inspector was ill informed. A little scary.
Swimming pool was tight! - nice looking backyard.
All workers are actors or in the biz, told us a story about Arnold voting here, one was a writer for the Tonight Show and told jokes
Polling place was in a kitchen, had to walk through whole mission, homeless guy yelled at us as we walked by

Are There Any Barriers to Voting in Polling Places?
A third set of questions check for the presence of several potential barriers that may inhibit equal access to the ballot. These questions are wide ranging, varying from assessing the number of poll workers and available machines to checking if voters are asked to show identification by poll workers or if provisional balloting is available. The number of poll workers varied from precinct to precinct, as did the number of booths or machines available to potential voters. Table 3 reports that most polling places had four or more poll workers (70%) and that virtually all of the new inkblot voting machines were working correctly. While we found that four out of five precincts used in 2004 had been used before (likely in 2002) as a polling place, 11% were reported to be new by the head poll worker and in 6% of the cases, the head poll worker was uncertain if the precinct had been used before. Of the roughly 1,700 precincts citywide, 187 were new locations, leaving about 150,000 registered voters potentially unfamiliar with their new polling place. When researchers asked the head poll worker if voters who were not listed on the rolls could vote a provisional ballot, more than 90% said yes, 3.5% said no, and 5% were uncertain. By state and federal law, any citizen who presents himself at a poll on Election Day has the opportunity to cast a provisional ballot if their name is not found on the voter rolls, and the provisional ballot is later checked at the county office to ensure that the citizen is a registered voter. To ensure that this option was available to all voters in California, Secretary of State Kevin Shelley required that the "Voter's Bill of Rights," which outlined this provision and others, be posted inside each polling place on Election Day. Despite this new requirement, 24% of all precincts did not have the voter rights posted inside the polling place. The Secretary of State of California instructs poll workers that no identification is necessary if the voter's name is correctly listed on the rolls. However, when researchers asked the head poll worker what form of identification voters needed to present before voting, a surprising 29% responded that some form of state identification was necessary. This may provide the single largest barrier to voting if poll workers at three in ten precincts in Los Angeles are asking voters to present identification.

[ Table 3 ]
Student observations:

It was a predominantly Armenian area, yet none of the workers spoke Armenian. They had to bring in one guy who spoke Armenian to have him translate to a man who didn't speak English.
The inspector would not let us examine the area because he said we were soliciting. He was beligerent and we could not get information.
Nice people, made sure we got everything we needed.
The precinct coordinator, did not let us interview the poll workers, stating that they don't have answer our questions. After showing him the voters bill of rights he directed us to a number we could call to answer our questions. We were not disrupting their duties, as no one was voting at the time.
Inspector hostile, trying to insist that voter's bill of rights was posted and it clearly wasn't. Trying to tell us we were disturbing voters. We were not.
Very nice ladies, they gave us "I voted" stickers. They were very helpful.
She wanted money for answering questions
Head poll worker said voters needed ID, fellow poll worker corrected him
Need a drivers license, military I.D. or utility bill to vote.
Asked for a California drivers license or a bill or a checking account.
Very friendly but said drivers license or voter registration card is required.
Did not know about provisional ballot
There was an older woman who was complaining about the government and the poll workers were trying to get her to leave
Woman being challenged/ Poll workers said people have been coming to wrong precinct all day. This poll is new
Poll worker said a state worker took out 2 voters cause they weren't on list and didn't want them to vote provisional.

IV. Income, Race, Precinct Quality and Voter Turnout

Precinct Differences by Income and Race
While the findings above demonstrated that not all precincts in Los Angeles are created equal, it is unclear whether these differences are randomly distributed throughout the city, or whether they are more likely to occur in certain parts of town. Using census tract data from the 2000 Census, we have overlaid each precinct with its corresponding census tract, and provided neighborhood level data on median income and racial/ethnic demographics. For income, we divide precincts into three equal segments, low, medium and high where precincts averaging a median household income of less than $30,913 are counted as low and those with median incomes of more than $49,226 counted as high. For race and ethnicity, we divided all precincts into four equal categories within each of the four groups included: White, Latino, Black and Asian, and compare results for only those precincts in the highest range for each of the racial/ethnic groups. For example, precincts in the quartile "lowest white" range from 0 to 3 percent white, while precincts in the quartile "highest white" range from 65 to 91 percent white. Appendix B displays the full range for each of the race categories .
Tables 4 - 6 report the full results of our precinct survey by income and race. Given that our sample yields a +/- 2 percent margin of error, between group differences of more than 5 percent can be considered statistically significant and are noted with an asterisk. Among the accessibility criteria (Table 4), high income and White precincts were more likely to have the precinct address posted in clear sight, and more likely to have adequate outside lighting to illuminate the precinct location at night. Interestingly, precincts in Latino neighborhoods were the most likely to have flags or banners indicating the presence of a polling place.

[ Table 4 ]
With regard to parking and other markers of precinct quality, many differences surfaced based on income and race (Table 5). Once again, high income and White precincts were the most likely to provide adequate parking nearby the precinct. While 75% of high-income precincts had adequate parking, just 61% of low-income precincts had accessible parking. Similarly, 72% of White precincts reported parking nearby compared to only 61% in Latino and 60% in Asian neighborhoods. The same trend existed for handicap accessibility with low income and non-White polling locations the least likely to have ramps and entrances handicap-ready. Inside, there was not a consistent pattern with respect to precinct size, however some points are noteworthy. Asian precincts were the least likely to be described as "very large" while Black precincts were the least likely to have waiting areas with chairs or sofas. Finally, although a majority of precincts did not provide additional amenities to voters such as coffee, donuts or pianists in the lobby, high income and White precincts were about twice as likely to do so.

[ Table 5 ]
Most troubling perhaps are the differences reported in Table 6 regarding barriers to voting. Once more, high income and White precincts are positively advantaged on issues related to ease of voting and voting rights. For example, high-income and White neighborhoods are the most likely to employ precincts that have been used in past elections. If voters become well acquainted with their voting place, it is reasonable to assume they will have fewer difficulties getting there on Election Day. However, if the polling place location is different in every election, voters must learn the name and address of the new precinct, and then find out where it is. When our student researchers asked the head poll worker about the need to present identification before voting, 35% in low-income precincts asked for I.D. compared to 22% in high income precincts. Similarly, about 21% of poll workers in White precincts required I.D. to vote, half the rate in Latino precincts where 39% of poll workers required I.D. to vote. Black and Asian precincts also reported higher rates of poll workers asking for I.D. than White precincts. Along the same lines, poll workers in high-income and White precincts were more likely to be aware of the provisional balloting rules and to allow voters not listed on the rolls to cast a provisional ballot. Interestingly, Latino precincts were the most likely, at 80%, to have the voter bill of rights posted, compared to Black precincts which were the least likely.

[ Table 6 ]

Precinct Differences and Voter Turnout
To assess the impact of precinct accessibility and quality on voter turnout, we matched aggregate Election Day turnout percentages for each precinct from the Los Angeles County Statement of Votes Cast by Precinct available on their website . The turnout rate purposely excludes absentee ballots and among the precincts in our study, the mean Election Day turnout was 27.60 percent and the median was 27.43 percent. It is important to keep in mind that most of the variance in the decision to vote is being accounted for by other individual level variables such as interest in politics, education, income, age and race that have been well documented in the voluminous scholarship on turnout. Here, we narrow our focus to the potential demobilizing costs associated with precinct accessibility and quality, while still controlling for some demographic factors. "While increasing the accessibility of precinct locations may not result in 30 or 40 point rise in participation, we would consider it a significant and policy relevant finding if the accessibility made the difference of between 0.3 and 1 point in the level of turnout," (Gimpel and Schuknecht 2003: 473). Thus, we use the same standard to evaluate means comparisons for turnout rates across precincts that may be more or less accessible.

Testing the effects of precinct quality on voter turnout without accounting for income and race, our two main control variables would be problematic. Indeed, the findings of the previous section reveal that precinct quality is correlated with income and race. Further, bivariate correlation analysis indicates that most of our independent measures of precinct quality and accessibility are highly collinear suggesting that a scale of precinct quality alongside demographic controls is the most appropriate way to test our hypothesis. In addition to a ten-point scale of precinct quality, we include one additional independent measure from our check list - whether or not the precinct had been used before, and label this measure precinct stability. Thus our regression equation for predicting voter turnout at the precinct level is:
Y = ß1(incomex) + ß2(latinox) + ß3(blackx) + ß4(asianx) + ß5(qualityx) + ß6(stabilityx) + constant

Our key independent variable, precinct quality, is constructed based on eight items on our check list and ranges from 0 - 10. The questions reflect many aspects of the precinct that a voter will encounter on Election Day.

These items are:

Were any signs, flags, or banners visible from outside such as "vote here"?
Was adequate parking space available adjacent to the polling place?
Was a restroom or restroom sign clearly marked inside the precinct?
Generally speaking, was the interior of the polling place well lit for reading?
Were there any additional amenities available to voters (such as coffee, donuts, snacks)?
Was there adequate outside lighting to make the precinct visible at night?
How many poll workers were present at the time of the visit?
Did all voting machines seem to be working properly?

The results of our regression are presented in Table 7 and uncover that, even controlling for income and race, precinct quality and precinct stability significantly effect voter turnout. Consistent with numerous studies on turnout, income has a positive effect and percent minority has a negative effect on aggregate turnout (Campbell et. al. 1960; Wolfinger 1980; Rosenstone and Hansen 1993; Verba, Schlozman and Brady 1995). As we anticipated, precinct quality does have a positive impact on voter turnout, and the substantive impact is rather large. The unstandardized coefficient (reported in column three) for precinct quality indicates that a one unit shift in the quality results in a 0.41 increase in voter turnout so that a ten unit shift in precinct quality from lowest to highest quality should result in a boost of 4.1 points to that precincts voter turnout rate. Similarly, a precinct that has been used before will see turnout about 1.5 points higher than a newly used precinct, all other things being equal. That our measure of precinct stability is significant, even when separately controlling for various components of precinct quality is remarkable, and speaks to the importance of regularly established precinct locations, so that voters can become familiar with their neighborhood polling place.

[ Table 7 ]
In order to depict these differences, we calculated the predicted voter turnout for precincts of varying qualities, while holding all other values at their mean. Figure 1 plots the predicted turnout rate for precincts at each step on the ten-point quality scale, as compared to the actual mean turnout level for precincts based on their quality score. Overall, the predicted regression line is a fairly accurate fit to the observed data. Next, figure 2 interacts precinct quality with precinct stability to generate the two extremes of our precinct study: (1) low-quality precincts being used for the first time; and (2) high-quality precincts that have been used in an election before. Precincts described as "lowest" quality generate an expected turnout rate of 23.2 percent, compared to 28.8 percent for "highest" quality precincts. This difference of 5.6 percentage points could easily alter the outcome close elections within the City of Los Angeles.

[ Figures 1 & 2 ]

Conclusion
Since the VRA was instituted, barriers to voting at the polling place have all but disappeared. However, irregularities in practices at some precincts in Florida during the 2000 election called into question whether new barriers to the ballot box exist. These barriers might include lack of signage or poor visibility, lack of adequate parking, no outside lighting, insufficient or poorly trained poll workers, or lack of stability in precinct location. Given that many citizens do not accurately assess the costs and benefits of voting, even small increases in the costs can lead to large decreases in voter turnout. This research has examined precinct accessibility and quality throughout Los Angeles and found that voters face a wide variation in the quality of precincts available to cast a ballot. More troubling, the findings here suggest that low quality precincts are not randomly distributed across the city and instead are more likely to be found in low-income and minority neighborhoods. While these communities already experience lower rates of voter turnout, the existence of many low quality precincts further depresses rates of voting, even after controlling for income and race. In this study we have highlighted several new costs of voting. However further research into this subject is needed to determine what other precinct-related barriers exist. In the meantime, policy makers and election officials in Los Angeles should take whatever measures are possible and practical to ensure that all polling places are equally visible, accessible, and user-friendly.

Tables

Table 1. Precinct Accessibility

Yes No Somewhat
Address was in clear sight 77.9 22.1
Adjacent to major street 60.6 39.4
Readily visible from street 69.1 11.0 20.0
Flags or banners made visible 75.8 24.2
Adequate outside lighting 76.5 23.5

In your opinion, how easy was the polling place to find?
Very easy 48.0
Somewhat easy 36.3
Somewhat difficult 12.2
Very difficult 3.6

Table 2. Precinct Quality

Yes No
Adequate parking nearby 67.3 32.7
Entrance handicap accessible 81.1 18.2
Restroom clearly marked 34.1 65.9
Interior well lit for reading 87.8 12.2
Additional amenities for voters 15.4 84.6

How small or large was the inside of the polling place?
Very small 17.7
Somewhat small 18.3
Medium 26.6
Somewhat large 18.7
Very large 18.7

What kind of waiting area was present?
None 16.7
Small standing area 33.4
Large standing area 21.5
Chairs/Sofas 28.4

Table 3. Barriers to Voting

Yes No Don't Know
Precinct used previously 82.4 11.2 6.4
Poll worker lives nearby 68.1 31.2 0.7
All voting machines working 96.6 3.4
Voter bill of rights posted 75.1 24.2 0.8
Four or more poll workers 70.0 30.0
Need to show ID to vote 29.2 70.2 0.6
Can vote provisional ballot 91.1 3.5 5.3

Table 4. Precinct Accessibility by Income and Race

Income Race/Ethnicity
Percent Reporting "Yes" Low Medium High White Latino Black Asian
Address was in clear sight 76.7 73.6 83.2 * 85.9 75.4 74.6 72.8 *
Adjacent to major street 62.9 63.3 55.6 * 59.0 61.5 62.2 59.3
Readily visible from street 70.6 68.3 68.6 68.4 67.4 70.3 70.5
Flags or banners made visible 76.7 68.3 73.5 70.8 78.9 74.3 64.8 *
Adequate outside lighting 74.3 71.9 83.3 * 80.1 74.0 68.3 73.8 *

In your opinion, how easy was the polling place to find?
Very easy 52.7 43.3 48.0 46.0 47.4 54.1 44.1 *
Somewhat easy 34.4 35.4 39.1 38.3 38.4 30.7 35.2 *
Somewhat difficult 11.7 15.1 9.8 11.5 12.1 12.6 14.6
Very difficult 1.3 6.3 3.2 4.3 2.2 2.6 6.1

* indicates difference of 5 percentage points or more
Race/Ethnicity categories refer to precincts that are in the top 25% of each race classification

Table 5. Precinct Quality by Income and Race

Income Race/Ethnicity
Percent Reporting "Yes" Low Medium High White Latino Black Asian
Adequate parking nearby 61.0 66.0 75.0 * 72.1 61.3 69.4 60.2 *
Entrance handicap accessible 76.2 81.9 87.3 * 86.1 74.3 78.5 81.2 *
Restroom clearly marked 40.5 31.3 30.8 * 26.1 40.3 37.3 31.9 *
Interior well lit for reading 88.4 85.8 89.4 87.5 86.5 88.4 88.4
Additional amenities for voters 11.5 16.0 18.6 * 19.4 10.3 13.8 13.7 *

How small or large was the inside of the polling place?
Very small 16.6 19.2 17.5 20.8 16.5 15.2 15.5 *
Somewhat small 16.2 17.6 21.0 18.2 17.8 17.4 23.7 *
Medium 26.0 30.4 23.3 22.9 24.8 29.5 27.1 *
Somewhat large 22.1 18.5 15.6 * 14.7 19.1 21.4 18.8 *
Very large 19.2 14.4 22.6 23.4 21.7 16.5 15.0 *

What kind of waiting area was present?
None 15.6 15.3 19.2 22.4 15.0 16.3 11.5 *
Small standing area 33.1 36.2 31.0 26.8 31.4 38.5 39.9 *
Large standing area 20.8 23.1 20.5 22.4 22.6 24.3 20.2
Chairs/Sofas 30.5 25.4 29.4 28.5 31.0 19.9 28.4 *

* indicates difference of 5 percentage points or more
Race/Ethnicity categories refer to precincts that are in the top 25% of each race classification

Table 6. Barriers to Voting by Income and Race

Income Race/Ethnicity
Percent Reporting "Yes" Low Medium High White Latino Black Asian
Precinct used previously 81.9 78.3 86.9 * 88.4 82.3 77.8 77.5 *
Poll worker lives nearby 74.6 63.8 66.1 * 65.8 73.0 75.9 67.1 *
All voting machines working 97.4 94.2 98.1 98.3 96.5 98.6 94.7
Voter bill of rights posted 77.7 74.4 73.4 73.7 80.4 69.1 78.6 *
Four or more poll workers 71.7 70.1 68.3 64.8 69.1 73.8 72.9 *
Need to show ID to vote 35.4 30.2 22.0 * 20.8 39.4 33.3 27.6 *
Can vote provisional ballot 87.3 93.7 92.3 * 94.4 88.9 88.4 90.0 *

* indicates difference of 5 percentage points or more
Race/Ethnicity categories refer to precincts that are in the top 25% of each race classification

Table 7. OLS Regression Results Predicting Precinct Level Voter Turnout

Variables Beta Coef.
Income .1912 .0001(.0000) ****
Percent Latino -.4799 -.1216(.0107) ***
Percent Black -.1031 -.0386(.0135) **
Percent Asian .0154 .0131(.0292)
Precinct Quality .0697 .0041(.0017) **
Precinct Stability .0718 .0150(.0062) **
Constant -- .2613(.0180) ***
N 679
F 77.69
Adj. R2 .4043
*** p < .001 ** p < .050

Figures
NOTE: Figures based on OLS unstandardized beta coefficients for equation:
Y = ß1(incomemedian) + ß2(latinomedian) + ß3(blackmedian) + ß4(asianmedian) + ß5(quality) + ß6(stability) + constant
References

Blais, André. 2000. To Vote or Not to Vote? The Merits and Limits of Rational Choice Theory.
Pittsburgh: University of Pittsburgh Press.

Brian, Craig and Bernard Grofman. 2001. "Election Day Registration's Effect on US Voter
Turnout." Social Science Quarterly. 82(1): 170-83.

Campbell, Angus, Philip E. Converse, Warren E. Miller, and Donald E. Stokes. 1960. The
American Voter. New York: John Wiley & Sons.

Downs, Anthony. 1957. An Economic Theory of Democracy. New York: Harper Collins.

Gimpel, J.G. and J.E. Schuknecht. 2003. "Political participation and the accessibility of the
ballot box." Political Geography. 22: 471-488.

Highton, Benjamin. 1997. "Easy Registration and Voter Turnout." Journal of Politics. 59(2):
565-575.

Navarro, Mireya and Somini Sengupta. 2000. "Arriving at Some Florida Voting Places,
Some Blacks Found Frustration." The New York Times. November 30. A1.

Katosh, John P. and Michael W. Traugott. 1982. "Costs and Values in the Calculus of
Voting." The American Journal of Political Science. 26(2):361-376

Knack, Stephen. 1995. "Does Rain Help the Republicans? Theory and Evidence on Turnout
and the Vote." Public Choice. 79(1):187-209.

Merrifield, John. 1993. "The Institutional and Political Factors that Influence Voter
Turnout." Public Choice. 77(3):657-669.

Riker, William and Peter Ordeshook. 1968. "A Theory of the Calculus of Voting." The
American Political Science Review. 62:25-42.

Rosenstone, Steven J. and John Mark Hansen. 1993. Mobilization, Participation, and Democracy
in America. New York: Macmillan Publishing.

Teixeira, Ruy A. 1987. Why Americans Don't Vote: Turnout Decline in the United States, 1960-
1984. New York: Greenwood Press.

Watternberg, Martin. 2002. Where Have All the Voters Gone? Cambridge: Harvard Press.

Wolfinger, Raymond E. and Steven J. Rosenstone. 1980. Who Votes? New Haven: Yale
University Press.

Verba, Sidney, Kay Lehman Schlozman, and Henry E. Brady. 1995. Voice and Equality:
Civic Volunteerism in American Politics. Cambridge: Harvard University Press.

Appendix A:
Center for the Study of Los Angeles
March 2, 2004 Polling Place Quality Check List

Team Name __________________________________________________________ Time of Day _____________

Precinct No. _____________________________ Camera Exposure No. _________________________________

Address _________________________________________________________________________________________

1) Was the address of the polling place in clear sight on the outside of the precinct? ? Yes ? No

2) Was the correct address given for the polling place? ? Yes ? No

3) How easy was the polling place to find? ? Very Easy ? Somewhat Easy ? Somewhat Difficult ? Very Difficult

4) Was the polling place adjacent to a major street (4 lanes/divided traffic)? ? Yes ? No

5) Was the polling place readily visible from the street? ? Yes ? Somewhat ? No

6) Were any signs, flags, or banners visible from outside such as "vote here"? ? Yes ? No

7) If Yes, did the sign make it obvious that this was a polling place? ? Yes ? No

8) Was adequate parking space available adjacent to the polling place? ? Yes ? No

9) Was the entrance to the polling place handicap (wheel chair) accessible? ? Yes ? No

10) Describe the general appearance and accessibility of the outside of the polling place. Provide any details you believe may effect citizens' access to the polling place or willingness to vote:

____________________________________________________________________________________________________

____________________________________________________________________________________________________

____________________________________________________________________________________________________

____________________________________________________________________________________________________

11) Generally speaking, how small or large was the inside of the polling place? ? Very Small ? Somewhat Small ? Medium
? Somewhat Large ? Very Large

12) What kind of waiting area was present? ? None ? Small Standing Area ? Large Standing Area ? Chairs/Sofas

13) Was a restroom or restroom sign clearly marked inside the precinct? ? Yes ? No

14) Generally speaking, was the interior of the polling place well lit for reading? ? Yes ? No

15) Were there any additional amenities available to voters (such as coffee, donuts, snacks)? ? Yes ? No

Explain _____________________________________________________________________________________________

____________________________________________________________________________________________________

16) Generally speaking, when you visited the precinct how safe did you feel? ? Very ? Somewhat ? Not Safe ? DK

17) Was there adequate outside lighting to make the precinct visible at night? ? Yes ? No
(NOTE: If you visit during daylight hours, inspect outside area for light fixtures, if necessary, ask this question of the head poll worker)

Center for the Study of Los Angeles
March 2, 2004 Polling Place Accessibility Study

( Page 2)

18) How many poll workers were present at the time of the visit? ? None ? One ? Two ? Three ? Four or More

19) Record the following demographic information for each poll worker:
A) Worker 1: ? Male ? Female ? White ? Black ? Latino ? Asian ? Under 40 ? 40 - 65 ? Over 65
B) Worker 2: ? Male ? Female ? White ? Black ? Latino ? Asian ? Under 40 ? 40 - 65 ? Over 65
C) Worker 3: ? Male ? Female ? White ? Black ? Latino ? Asian ? Under 40 ? 40 - 65 ? Over 65
D) Worker 4: ? Male ? Female ? White ? Black ? Latino ? Asian ? Under 40 ? 40 - 65 ? Over 65
E) Worker 5: ? Male ? Female ? White ? Black ? Latino ? Asian ? Under 40 ? 40 - 65 ? Over 65
F) Worker 6: ? Male ? Female ? White ? Black ? Latino ? Asian ? Under 40 ? 40 - 65 ? Over 65

20) How many polling booths/voting machines were available to voters? _________

21) Did all voting machines seem to be working properly? ? Yes ? No

22) About how many people were waiting in line to vote when you visited? ? Less than 10 ? 11-25 ? 26-50 ? Over 50

23) Ask the head poll worker, "Do you live here in this polling place jurisdiction"? ? Yes ? No ? DK

24) Ask the head poll worker, "Was this polling place used in a previous election"? ? Yes ? No ? DK

25) Ask the head poll worker, "Other than English, what other languages do poll workers speak: (list all)

____________________________________________________________________________________________________

26) Ask the head poll worker, "If I live in this neighborhood, what form of ID do I need to vote here when I check in?"

? No ID needed ? Some ID is needed ? Poll worker doesn't know

27) Ask the head poll worker, "If somebody shows up to vote, but for some reason they are not listed on the voter directory, can they still vote or do they need to be listed?"
? Yes they can vote ? No, need to be listed ? Depends/Maybe ? DK

28) Was the "Voter's Bill of Rights" clearly posted inside the polling precinct? ? Yes ? No ? DK

29) Additional observations about the polling place: (be as specific as possible)

____________________________________________________________________________________________________

____________________________________________________________________________________________________

____________________________________________________________________________________________________

____________________________________________________________________________________________________

____________________________________________________________________________________________________

____________________________________________________________________________________________________

If you have any questions during the day please call the Center for the Study of Los Angeles
Phone: (310) 338 - 4565

Appendix B:
Distribution of Race and Ethnicity Quartiles

White Latino Black Asian
Quartile 1 .00 - .03 .01 - .11 .00 - .02 .00 - .02
"lowest"
Quartile 2 .04 - .29 .12 - .35 .03 - .04 .03 - .06
"med low"
Quartile 3 .30 - .64 .36 - .64 .05 - .13 .07 - .12
"med high"
Quartile 4 .65 - .91 .65 - .99 .15 - .92 .13 - .56
"highest"

Special Thanks to the Undergraduate Research Team:

Abby Onderdonk
Adrian Canzoneri
Alberto Ramirez
Alex Aguila
Alissa McGlover
Amber Nevada
Amy Ashbaugh
Angel Bocha
Antonio Plascencia, Jr.
Aram Nadjarian
Ashley Moorhead
Audra Shackelford
Casandra
Celica Macias
Chereese Waters
Christine Thorson
Cristine Carey
Daisy Beltran
Daisy Trudeau
Danny Cup Choy
Denise
Dominic Quiller
Edythe Huang
Elizabeth Zamudio
Erika B. Hallum
Ester Seflyan
Hayon Abdulhafiz
Hilary Yribarien
Iris Patton
Ivan Castillo
Ivette Nuñez
Jamie Steiner
Jeff Mckirdy
Jessica Sach
Johanna Lopez
Juanita Guillen
Judy Hernandez
Juliana Moran
Karla Arvizu
Karla Solano
Kelseve Olson
Kim Hardina
Kristi Hg
Kyle Weinheimer
Laura Jeffords
Lindsay Sargent
Lis Chong
Luis Valdivia
Manar Fakhoury
Marc Santos
Marco Villacorta
Maria Benitez
Maria E. Jauregui
Marina Gonzalez
Marthew Pomeroy
Marvin Benitez
Maryland Sharidan
Matt Delja
Matt Dino
Matt Sherwood
Meghan Moore
Melissa Banda
Meredith McCall
Michael Recker
Michael W. Russell
Milagro Romero
Miriam Ali
Nasise Tobin
Ofelia Medina
Pat Furlong
Paul Schroeder
Rafael Reyes
Ramon Guerra
Rick
Rosa Ramirez
Rose Thomas
Rose Vasilj
S. Elington
Samira Bordji
Sean Mumper
Shari Pourebrahim
Stephanie Jensen
Sylvia Ciborawski
Taryn Vogel
Tim Murphy
Tobin Nasise
Tracy Cruz
Troy Matthews
Vanessa Elfend
Vanessa Quiñones
Veronica Diaz
Veronica Sanchez
Victoria Lau
Whitney Izzo