A healthy democracy requires that every voter has equal access to an effective vote. Every voter should be able to easily express their preferences and have confidence that every effort has been made to record and count those preferences accurately. But poor ballot design makes it harder for people to vote.
Hundreds of thousands of votes go uncounted every election season because of poorly designed ballots, and numerous studies have found evidence that counties with higher concentrations of Black, Hispanic, low-income, less educated, and/or older voters have higher rates of unrecorded votes due, in part, to bad ballot design.
Adhering to straightforward, science-based ballot design recommendations can effectively lower ballot rejection rates, decrease the number of ballot envelopes that need to be cured or fixed, decrease the time it takes to vote and increase voter turnout.
Ballot design is a crucial yet often overlooked part of election administration in the United States. To ascertain the general usability of ballots across the nation, we collected and evaluated sample ballots from one jurisdiction in each of eight counties in five traditionally electorally competitive and significant states—Arizona, Georgia, North Carolina, Ohio, and Pennsylvania. We evaluated the design of these sample ballots using 10 of our equitable ballot design recommendations.
Evaluating Ballot Design in the 2024 General Election
Introduction
A healthy democracy requires that every voter has access to an equal and effective vote. Every voter should be able to easily express their preferences and have confidence that every effort has been made to record and count those preferences accurately. But poor ballot design makes it harder for people to vote. Hundreds of thousands of votes go uncounted every election season because of poorly designed ballots (Norden and Iyer 2011; Norden et al. 2008; Norden, Quesenbery, and Kimball 2012). Furthermore, numerous studies have found evidence that counties with higher concentrations of Black, Hispanic, low-income, less educated, and/or older voters have higher rates of unrecorded votes due, in part, to bad ballot design (Darcy and Schneider 1989; Gordon-Rogers, Latner, and Williams 2024a; Herron and Sekhon 2003; Kimball and Kropf 2005; Kimball, Owens, and Keeney 2004; Knack and Kropf 2003; Norden and Iyer 2011; Norden, Quesenbery, and Kimball 2012; Sinclair and Alvarez 2004; Tomz and Van Houweling 2003).
Adhering to straightforward, science-based ballot design recommendations can effectively lower ballot rejection rates, decrease the number of ballot envelopes that need to be cured or fixed of errors to be counted (Norden, Quesenbery, and Kimball 2012), decrease the time it takes to vote (Engstrom and Roberts 2020; Spencer and Markovits 2010), and increase voter turnout (Fraga and Merseth 2016; Hopkins 2011; Jones-Correa 2005). For example, in 2025, Pennsylvania Secretary of State Al Schmidt said the Department of State's mail ballot envelope redesigns were one of the reasons the state saw lower ballot rejections in the 2024 general election (Scolforo 2025). [^1]
Ballot design is a crucial yet often overlooked part of election administration in the United States. To ascertain the general usability of ballots across the nation, we collected and evaluated sample ballots from one jurisdiction in each of eight counties [[2]] in five traditionally electorally competitive and significant states---Arizona, Georgia, North Carolina, Ohio, and Pennsylvania.
Evaluating Sample Ballots
The selected counties are generally the most populous counties in battleground states that have been crucial in determining the outcome of the Electoral College in recent presidential elections (Gordon-Rogers, Latner, and Williams 2024a). Also included are some smaller rural counties to account for potential localized differences in ballot design.
Using 10 of our equitable ballot design recommendations (Gordon-Rogers, Latner, and Williams 2024b), we evaluated the design of the sample ballots. It should be noted that the selected recommendations are not exhaustive and do not represent all the elements on which ballot design can and should be evaluated. The selection criteria for the scale of 10 recommendations were twofold. First, they are recommendations related to in-person paper ballots. Second, they are recommendations related to design features that are observable by examining images of sample ballots. Importantly, we do not contend that these recommendations encompass all of the most impactful or significant of our equitable ballot design recommendations. Consequently, a ballot's relative scores do not fully capture overall ballot design and usability but rather are representative of only the following recommendations.
Selected Design Recommendations Used in Analysis
1. Ballot instructions are horizontal.
2. Instructions are in the top left corner of the ballot.
Voters are more likely to unintentionally skip contests when the instructions are not clearly differentiated from the ballot contests. Instead of being formatted vertically, like contests are, we recommend that instructions are positioned horizontally in the top left corner of ballots, as voters usually look there first (Kimball and Kropf 2005). Second, when instructional text is placed in the upper left corner, "they appear as shorter lines of text, which are easier to read and comprehend than long lines spread over an entire page" (Kimball and Kropf 2005, 513; see also Christian and Dillman 2004; Dillman 2000).
3. Ballot has instructional illustrations.
The inclusion of instructional illustrations assists not only low-literacy voters but all voters (CCD n.d.). Illustrations should highlight the most important instructions and, of course, be accurate (CCD n.d.). Figure 1 (taken from Allegheny County's 2024 sample ballot) shows a common instructional illustration of how voters should mark contest choices. Figure 2 (taken from the same sample ballot) is a common illustration of how voters add a write-in candidate to the ballot.
4. Instructions appear immediately before associated tasks.
Various studies have found that instructions should appear directly before the associated tasks (Christian and Dillman 2004; Dillman 2000; Kimball and Kropf 2005). The law of proximity, a psychological theory, argues that placing items in proximity encourages readers to see them as related (Dillman 2000; Kimball and Kropf 2005). As a result, voters are more likely to fill out their ballots correctly.
5. Instructions include information on write-in candidates, voter errors, and the consequences of a spoiled ballot.
Instructions for a ballot must contain the information necessary for voters to effectively fill it out, which will consequently reduce voter errors. Voters should be informed about how to write in candidates, how to address or fix errors, and what happens with a spoiled ballot (CCD 2020; Norden, Quesenbery, and Kimball 2012).
6. Instructions use clear, simple language.
At times, the language used in ballot instructions can be overly technical, which can lead to voter confusion and errors. In Cuyahoga County, Ohio (one county examined in this analysis), included in their 2010 ballot instructions was the phrase "select the set of joint candidates." This confusing wording resulted in over 2,000 uncounted gubernatorial overvotes (Norden, Quesenbery, and Kimball 2012).
Research has shown that high levels of instruction comprehension [3] are facilitated when all written content is written at or below the eighth-grade reading level (Bryant et al. 2023; Reilly and Richey 2009). Consequently, we operationalize "simple, clear language" as instructions written equal to or lower than the eighth-grade reading level. The Flesch-Kincaid scale estimates the US grade level required to comprehend written text by analyzing the number of words per sentence and the number of syllables per word (Thomas, Hartley, and Kincaid 1975). This tool is commonly used across multiple fields (Ley and Florio 1996; Neuhoff et al. 2016). Due to occasional discrepancies found in computer-generated reading levels (Neuhoff et al. 2016), I take the mean of two scores, the first estimated with the built-in calculator in Microsoft Word (n.d.) and the second determined with an online calculator available on ReadabilityFormulas.com (n.d.). [4]
7. Candidate names for each race are in one continuous column.
Contests that divide candidate lists between two columns or pages have been shown to increase the number of undervotes (Norden et al. 2008) and/or overvotes (Cottrell et al. 2022; Kimball and Kropf 2017; Norden, Quesenbery, and Kimball 2012; Quesenbery 2016) because voters erroneously assume that the split contest is two separate races. A ballot with a split contest in the 2020 general election of Gwinnett County, Georgia, for example, resulted in about 4,200 overvotes in a senatorial special election contest (Cottrell et al. 2022).
8. Ballot uses color and contrast (e.g., shading, bolding, headings) to clearly demarcate races.
Shading and bolding are design choices often left to the discretion of local election officials. This results in substantial geographic variance (Kimball and Kropf 2005). When paper ballots do not effectively differentiate between sections, voters are more likely to unintentionally skip races and undervote (Bryant et al. 2023; Jefferson 2007; Norden, Quesenbery, and Kimball 2012). Employing shading, contrast, and color in a consistent manner can support ballot navigation and guide voters from one task to another (CCD n.d.; Dillman 2000; Kimball and Kropf 2005).
9. Ballot avoids using all capital letters.
Lowercase letters are easier to read than all capital letters because, as the Election Assistance Commission states, "They are easier to recognize" (EAC 2022, sec. 1; see also CCD n.d.). According to Lausen (2007), lowercase letters have unique silhouettes with a mix of ascenders (strokes that extend above the standard height of lowercase letters) and descenders (strokes that extend below the baseline). The combination of uppercase and lowercase letters gives readers visual cues to help them read more quickly and to more easily identify words (Lausen 2007).
10. Ballot avoids centered type.
Left-aligned type is easier to read than centered type, which forces the eye to stop reading to locate the beginning of the next line (EAC 2022; CCD n.d.). Centered type "is complex, decorative, and more demanding to read" (Lausen 2007, 31). Moreover, the placement of the beginning of each centered line is inconsistent. Ballots should apply the consistency of left alignment since our eyes naturally go from left to right (Lausen 2007).
Analysis
As shown in Table 1, the lowest-rated sample ballot is from Maricopa County, Arizona, receiving 5 out of 10. The highest-rated sample ballots received 7 out of 10 and are from Cuyahoga County, Ohio, and Allegheny County, Pennsylvania. The mode rating is 6 out of 10, with six of the eight counties examined receiving this rating. The mean rating is 6, with the sample standard deviation of 0.53.
All the sample ballots fail to adhere to the recommendations to position instructions in the top left corner of the ballot and to avoid all capital letters, with every ballot showing at least one instance of doing so. Each sample ballot uses contrast and color and positions instructions just before the associated tasks. Twenty-five percent of the samples, or two counties (Cuyahoga and Lorain), follow the recommendation of avoiding centered type.
All but one, Maricopa County, list candidate names in one continuous column as recommended. All but one, Lorain County, [5] include instructional illustrations. Instructions in three counties, Allegheny, Cuyahoga, and Lorain, include information on write-in candidates, voter mistakes, and spoiled ballots.
Five of the eight ballots, or 62.5 percent of the samples, display horizontal instructions, as recommended. The average Flesch-Kincaid combined reading level is 6.5. The highest reading level is 8 in Allegheny County, while the lowest is Maricopa County at 4.1.
Maricopa County, Arizona [6]
Because Maricopa County's sample ballots for the 2024 general election were available online only after providing voter information, we requested one through a public records request (MCE 2024). On a 10-point scale, the 2024 general election ballot is coded as 5. Sections are differentiated with contrast and color. Contest headers/subheaders, candidate names, and issue options are in all capital letters. In addition, contest headers/subheaders are center aligned. The ballot meets three of the four instruction-related criteria---being written in simple, clear language (4.1 average reading level), including illustrations, and being horizontal. As Figure 3 exhibits, despite adhering to these recommendations, the instructions are not in the left corner of the ballot and fail to provide information on fixing errors and spoiled ballots. Apart from our recommendations, the instructions are also in a relatively small type and have no bolding, shading, or color to draw the voter's attention.
Finally, there is a split contest. In the section for the retention of the judges of the superior court, the candidates are listed in two separate columns. Notably, there are 42 judges to be voted on. Although ballot length is not an evaluated criteria, as administrators have little to no control over the number of contests or issues, it is important to note that ballot length increases the likelihood of voting errors (Bernardo, Pearson-Merkowitz, and Macht 2022). [7]
Fulton County, Georgia
Fulton County's sample ballots for the 2024 general election were available on their election website (FC 2024). Out of 10 points, the sample ballot is coded as 6. Instructions, as shown in Figure 4, appear before associated tasks, are horizontal, include illustrations, and are written in clear, simple language (6.6 average reading level). The instructions do tell voters to not try to fix mistakes, but they do not explicitly inform voters that spoiled ballots will not be counted.
There are several cases of all-capital text in the instructions, which are written across the entire width of the ballot instead of in the upper left corner. The contest headers use centered-aligned text instead of aligning the text at the left margin, as recommended. There is shading to differentiate between sections, but it should be noted there is no shading to differentiate between contests. Meeting recommendations, there are no split contests; candidates are listed in continuous columns.
Columbus, Durham, and Mecklenburg Counties, North Carolina
As sample ballots in North Carolina were available to those with valid voter information, we obtained Columbus, Durham, and Mecklenburg Counties' sample ballots from a local election administrator (NCSBE n.d.). Since all three ballots have the same design, we discuss them here together.
On a 10-point scale, the 2024 general election sample ballots are coded as 6. Despite meeting four of the six instructions-related criteria (horizontal, positioned before associated tasks, having illustrational images, and written in clear, simple language with an average grade-level score of 6.2), the instructions do not appear in the top left corner or include information about the consequences of a spoiled ballot (though they do inform voters about how to write in candidates and fix errors). As shown in Figure 5, ballot section headers are in all capital letters, and headers and subheaders are left aligned. The ballot uses contrast and color to differentiate between sections. Finally, there are no split contests, and candidates are listed in one continuous column.
Cuyahoga County, Ohio
Cuyahoga County's sample ballots for the 2024 general election were available, so we requested one from a local election administrator (CCBOE n.d.). On a 10-point scale, the ballot is coded as 7. Figure 6 demonstrates that while the ballot's instructions are written in clear and simple language (7.2 average reading level), appear before associated tasks, and include accurate illustrations as well as information on how to write in candidates, fix errors, and deal with spoiled ballots, they are arranged in a vertical column, which is not the recommendation (horizontal and placement in the upper left). [8] The ballot uses shading to differentiate between races. The text is left aligned, though there is one instance of all capital letters. [9] There are no split contests, and candidate names are listed in continuous columns.
Lorain County, Ohio
Lorain County's sample ballots for the 2024 general election were available online (OHBOE 2024). Out of 10 points, the sample ballot is coded as 6. As seen in Figure 7, the instructions are not horizontal or in the upper left corner [10] and instead are vertical. Moreover, the instructions do not include any illustrations highlighting key information. The instructional text does use clear and simple language (7.2 average reading level), appears directly before associated tasks, and informs voters about write-in candidates, fixing errors, and the consequences of spoiled ballots.
Although the ballot generally avoids using all capital letters, there are at least three cases where they do appear. [11] No text on the ballot is centered aligned but instead is aligned at the left. There are no split contests, and candidates are listed in continuous columns. Finally, contests are differentiated with shading and bolding.
Allegheny County, Pennsylvania
Allegheny County's sample ballots for the 2024 general election were available on their election website (Allegheny County n.d.). On a 10-point scale, the sample ballot is coded as 6. As shown in Figure 8, the instructions appear before associated tasks, use clear, simple language (8.0 reading level), include instructional illustrations, and share information on how to write in candidates and fix errors and the consequences of a spoiled ballot, but they are vertical and not in the upper left corner of the ballot. [12] Additionally, there is one instance of all capital letters. The contest headers are center aligned, but they are properly differentiated, including by shading. Finally, there are no split contests, and candidates are listed in one continuous column.
Conclusion
Our analysis demonstrates there is room for improvement across every jurisdiction evaluated. No county's sample ballot received a perfect score. The highest-scoring sample ballot followed only 70 percent of the selected recommendations. Still, we recognize the complex nature of election administration and the numerous challenges administrators face in designing ballots, such as heeding instructions set by statute and including features required by voting technology. For example, the use of all capital letters, centering, and the wording of instructions is sometimes established by legal statute and is not subject to election official discretion.
Despite these barriers, election administrators are often able to make ballot design decisions (Engstrom and Roberts 2020) that make voting more accessible and equitable for all voters, including by adopting the recommendations discussed here as well as the extended recommendations discussed in Gordon-Rogers, Latner, and Williams (2024a). Additional analysis evaluating these and other ballots could highlight more modifications that further improve accessibility and equity. For example, as the number of voters who vote by mail increases, the design of absentee ballots is becoming more important. Examining absentee ballot instructions could yield important findings and design suggestions for election officials and administrators. Despite the popularity of voting by mail, President Trump has called for its reduction, and Utah, one of eight states that currently have universal vote-by-mail elections, is close to implementing a new law that would put an end to the current practice of automatically mailing ballots to registered voters (Marley and Sanchez 2025).
Separately, analyses focusing on language and disability accessibility design are critical to the mission of achieving more equitable voting. For example, researchers should analyze the formatting and accuracy of non-English language instructions on ballots with the help of local grassroots organizations that understand the unique regional dialects of a jurisdiction's populations.
Ensuring every voter has equal and effective access to the voting process through ballot design is essential to the health of our democracy. Poorly designed ballots not only hamper voters' ability to cast their vote accurately, they also disproportionately affect historically marginalized groups, such as Black and Hispanic voters. Our science-based recommendations can reduce barriers to accessible and equitable elections by lowering rejection rates, reducing the number of ballots in need of curing, and increasing turnout.
Author
Liza Gordon-Rogers is a research officer in the Center for Science and Democracy at the Union of Concerned Scientists.
Acknowledgments
This report was made possible through the generous support of the Bernard F. and Alva B. Gimbel Foundation, Democracy Fund, and UCS members.
We extend our sincere thanks to the UCS staff who contributed to this project and provided valuable feedback.
We are especially grateful to the members of our Election Science Task Force---Andrea Benjamin, Kamaryn Norris, Jennifer Morrell, Alec Ramsay, Scott Seeborg, and Quentin Turner---for their expertise and critical input in shaping and refining our recommendations.
We appreciate the contributions of our peer reviewers, Ivy Cargile; Michael Latner, director of Research on Democratic Reform at the Charles Hamilton Houston Institute for Race & Justice at Harvard Law School, professor of political science at California State Polytechnic University San Luis Obispo, and former UCS senior fellow; and Whitney Quesenbery, executive director of the Center for Civic Design.
Finally, we thank the organizations and representatives who assisted us at various stages of this process, including All Voting is Local, All Voting is Local Michigan, All Voting is Local Pennsylvania, Chelsea Jones and Kevin Morris of the Brennan Center for Justice, the Center for Civic Design, Common Cause, Common Cause Michigan, Detroit Action, Fair Count, Greater Cleveland Congregations, League of Conservation Voters Michigan, League of Women Voters of Michigan, Make the Road PA, Michigan AFL-CIO, Michigan United, Ohio Organizing Collaborative, Pennsylvania Voice, Progress Michigan, and Promote the Vote.
Organizational affiliations are listed for identification purposes only. The opinions expressed herein do not necessarily reflect those of the organizations that funded the work or the individuals who reviewed it. The Union of Concerned Scientists bears sole responsibility for the report's contents.
References
Allegheny County. n.d. "View Your Sample Ballot." Accessed January 28, 2025. https://apps.alleghenycounty.us/website/ElectPollPlace.aspx
Ballotpedia. n.d. "Readability Laws for Ballot Measure Language." Accessed February 5, 2025. https://ballotpedia.org/Readability_laws_for_ballot_measure_language
Bernardo, Nicholas D., Jr., Shanna Pearson-Merkowitz, and Gretchen A. Macht. 2022. "The Effect of Ballot Characteristics on the Likelihood of Voting Errors." State Politics & Policy Quarterly 22 (2): 121--39. https://doi.org/10.1017/spq.2021.24
Bryant, Lisa A., David Kimball, Gretchen Macht, Anita Manion, Mindy Romero, and Robert M. Stein. 2023. In-Person Voting: Best Practices and New Areas for Research. Cambridge, MA: MIT Election Data and Science Lab. https://electionlab.mit.edu/sites/default/files/2023-10/in-person-voting_MEAES.pdf
CCBOE (Cuyahoga County Board of Elections). n.d. "Get Your Voting Information." Accessed January 28, 2025. https://boe.cuyahogacounty.gov/voters/get-your-voting-information
CCD (Center for Civic Design). 2020. "Questions to Ask While Designing a Ballot: A Checklist for Usable Ballots." Last modified January 28, 2020. https://civicdesign.org/questions-to-ask-while-designing-a-ballot/
---------. n.d. "Designing Usable Ballots." Accessed January 27, 2025. https://civicdesign.org/fieldguides/designing-usable-ballots/
Christian, Leah Melani, and Don A. Dillman. 2004. "The Influence of Graphical and Symbolic Language Manipulations on Responses to Self-Administered Questions." Public Opinion Quarterly 68 (1): 57--80. https://doi.org/10.1093/poq/nfh004
Cottrell, David, Felix E. Herron, Michael C. Herron, and Daniel A. Smith. 2022. "Auditing the 2020 General Election in Georgia: Residual Vote Rates and a Confusing Ballot Format." Election Law Journal: Rules, Politics, and Policy 21 (1): 1--18. https://doi.org/10.1089/elj.2020.0708
Darcy, R., and Anne Schneider. 1989. "Confusing Ballots, Roll-Off, and the Black Vote." Political Research Quarterly 42 (3): 347--64. https://doi.org/10.1177/106591298904200309
Dillman, Don A. 2000. Mail and Internet Surveys: The Tailored Design Method. 2nd ed. New York: Wiley.
EAC (US Election Assistance Commission). 2022. "Designing Polling Place Materials." Last modified March 11, 2022. https://www.eac.gov/election-officials/designing-polling-place-materials
Eltohamy, Farah. 2020. "2020 Election Changes Include New Equipment for Maricopa County." Cronkite News (blog). February 27. https://cronkitenews.azpbs.org/2020/02/27/maricopa-county-voters-equipment/
Engstrom, Erik J., and Jason M. Roberts. 2020. The Politics of Ballot Design: How States Shape American Democracy. Cambridge: Cambridge University Press.
FC (Fulton County). 2024. "Composite Sample Ballot." Accessed January 28, 2025. https://www.fultoncountyga.gov/-/media/Departments/Registration-and-Elections/Sample-Ballot/2024-Sample-Ballots/November-5-General-Election/Composite-Sample-Ballot--November-5-2024-General-Electionrevised-9232024.pdf
Fifield, Jen. 2024. "Maricopa County's Two-Page Ballot Has Already Caused Some Problems ---And More May Be Coming." Votebeat, October 31, 2024. https://www.votebeat.org/arizona/2024/10/31/maricopa-county-encounters-problems-two-page-ballot/
Fraga, Bernard L., and Julie Lee Merseth. 2016. "Examining the Causal Impact of the Voting Rights Act Language Minority Provisions." Journal of Race, Ethnicity, and Politics 1 (1): 31--59. https://doi.org/10.1017/rep.2015.1
Gordon-Rogers, Liza, Michael Latner, and Christopher Williams. 2024a. Race and Representation in Battleground Counties: An Analysis of Racial Disparities in Voter Turnout and Ballot Rejections in the 2020 Presidential Election. Cambridge, MA: Union of Concerned Scientists. https://doi.org/10.47923/2024.15576
---------. 2024b. * Equitable Ballot Design and Voter Education Materials*. Cambridge, MA: Union of Concerned Scientists. https://doi.org/10.47923/2024.15577
Herron, M. C., and J. S. Sekhon. 2003. "Overvoting and Representation: An Examination of Overvoted Presidential Ballots in Broward and Miami-Dade Counties." Electoral Studies 22 (1): 21--47. https://doi.org/10.1016/S0261-3794(01)00052-X
Hopkins, Daniel J. 2011. "Translating into Votes: The Electoral Impacts of Spanish-Language Ballots." American Journal of Political Science 55 (4): 814--30. https://doi.org/10.1111/j.1540-5907.2011.00534.x
Jefferson, David. 2007. "What Happened in Sarasota County?" The Bridge: Linking Engineering and Society 37 (2). https://nae.edu/7665/WhatHappenedinSarasotaCounty
Jones-Correa, Michael. 2005. "Language Provisions under the Voting Rights Act: How Effective Are They?" Social Science Quarterly 86 (3): 549--64. https://doi.org/10.1111/j.0038-4941.2005.00317.x
Kimball, David C., and Martha Kropf. 2005. "Ballot Design and Unrecorded Votes on Paper-Based Ballots." Public Opinion Quarterly 69 (4): 508--29. https://doi.org/10.1093/poq/nfi054
---------. 2017. "Analysis of Overvotes in the 2016 California Senate Primary Election" (working paper, Center for Civic Design, Cambridge, MA). https://www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=https://civicdesign.org/wp-content/uploads/2016/04/Analysis-of-Overvotes-2016-Senatt-Kimball-Kropf.docx&ved=2ahUKEwjn-Ki-ys2MAxU5FVkFHZPeOW0QFnoECAoQAQ&usg=AOvVaw3RKb-LF55i168tAsCHHwaf
Kimball, David C., Chris T. Owens, and Katherine McAndrew Keeney. 2004. "Unrecorded Votes and Political Representation." In Counting Votes: Lessons from the 2000 Presidential Election in Florida, edited by Robert P. Watson, 135--50. Gainesville: University of Florida Press.
Knack, Stephen, and Martha Kropf. 2003. "Voided Ballots in the 1996 Presidential Election: A County-Level Analysis." Journal of Politics 65 (3): 881--97. https://doi.org/10.1111/1468-2508.00217
Lausen, Marcia. 2007. Design for Democracy: Ballot and Election Design. Chicago: University of Chicago Press. https://press.uchicago.edu/ucp/books/book/chicago/D/bo5723477.html
Ley, P., and T. Florio. 1996. "The Use of Readability Formulas in Health Care." Psychology, Health & Medicine 1 (1): 7--28. https://doi.org/10.1080/13548509608400003
Marley, Patrick, and Yvonne Wingett Sanchez. 2025. "Utah on Path to End Automatic Mail Ballots after Wave of Misinformation." Washington Post, March 17, 2025. https://www.washingtonpost.com/politics/2025/03/17/utah-mail-voting-election-misinformation/
MCE (Maricopa County Elections). 2024. "2024 General Election Information Now Available Online." Press release, September 16. https://elections.maricopa.gov/news-and-information/elections-news/2024-General--Election-Information-Now-Available-Online.html
Microsoft. n.d. "Get Your Document's Readability and Level Statistics." Accessed February 2, 2025. https://support.microsoft.com/en-us/office/get-your-document-s-readability-and-level-statistics-85b4969e-e80a-4777-8dd3-f7fc3c8b3fd2
NCSBE (North Carolina State Board of Elections). n.d. Voter Search: County; accessed January 29, 2025. https://vt.ncsbe.gov/RegLkup/
Neuhoff, Emily, Kristiana Feeser, Kayla Sutherland, and Thomas Hovatter. 2016. "Flesch-Kincaid Reading Grade Level Re-examined: Creating a Uniform Method for Calculating Readability on a Certification Exam." Online Journal for Workforce Education and Development 9 (1), Article 2. https://opensiuc.lib.siu.edu/ojwed/vol9/iss1/2
Norden, Lawrence, and Sundeep Iyer. 2011. Design Deficiencies and Lost Votes. New York: Brennan Center for Justice. https://www.brennancenter.org/our-work/research-reports/design-deficiencies-and-lost-votes
Norden, Lawrence, David Kimball, Whitney Quesenbery, and Margaret Chen. 2008. Better Ballots. New York: Brennan Center for Justice. https://www.brennancenter.org/our-work/research-reports/better-ballots
Norden, Lawrence, Whitney Quesenbery, and David C. Kimball. 2012. Better Design, Better Elections. New York: Brennan Center for Justice. https://www.brennancenter.org/our-work/research-reports/better-design-better-elections
OHBOE (Ohio Bureau of Elections). 2025. View Ballots: 11/5/2024 General Election; accessed January 28, 2025. https://lookup.boe.ohio.gov/vtrapp/lorain/ballotlist.aspx
Quesenbery, Whitney. 2016. "Breaking the Ballot: 34 Candidates for Senate." Center for Civic Design. News release, April 5. https://civicdesign.org/breaking-the-ballot/
ReadabilityFormulas.com. n.d. "Readability Scoring System PLUS." Accessed February 4, 2025. https://readabilityformulas.com/readability-scoring-system.php
Reilly, Shauna, and Sean Richey. 2009. "Ballot Question Readability and Roll-Off: The Impact of Language Complexity." Political Research Quarterly 64 (1): 59--67. https://doi.org/10.1177/1065912909349629
Scolforo, Mark. 2025. "Pennsylvania Elections Chief Touts Progress in Reducing Mail Ballot Rejection Rate." Associated Press, January 31, 2025. https://apnews.com/article/pennsylvania-mail-ballots-election-voting-f51bfba8910686eb0f63c15fb6757f20
Sinclair, D. E. "Betsy," and R. Michael Alvarez. 2004. "Who Overvotes, Who Undervotes, Using Punchcards? Evidence from Los Angeles County." Political Research Quarterly 57 (1): 15--25. https://doi.org/10.1177/106591290405700102
Spencer, Douglas M., and Zachary S. Markovits. 2010. "Long Lines at Polling Stations? Observations from an Election Day Field Study." Election Law Journal: Rules, Politics, and Policy 9 (1): 3--17. https://doi.org/10.1089/elj.2009.0046
Thomas, Georgelle, R. Derald Hartley, and J. Peter Kincaid. 1975. "Test-Retest and Inter-Analyst Reliability of the Automated Readability Index, Flesch Reading Ease Score, and the Fog Count." Journal of Reading Behavior 7 (2): 149--54. https://doi.org/10.1080/10862967509547131
Tomz, Michael, and Robert P. Van Houweling. 2003. "How Does Voting Equipment Affect the Racial Gap in Voided Ballots?" American Journal of Political Science 47 (1): 46--60. https://doi.org/10.1111/1540-5907.00004
Ulrich, Steve. 2025. "Rejected Mail Ballots in PA Fall by 57% from 2024 Spring Primary." Politics PA, January 24, 2025. https://www.politicspa.com/rejected-mail-ballots-in-pa-fall-by-57-from-2024-spring-primary/140531/
Footnotes
1. There was a 57 percent decrease in rejected mail ballots in the general election from the 2024 primary election (Ulrich 2025).
2. The counties are Maricopa County, Arizona; Fulton County, Georgia; Columbus, Durham, and Mecklenburg Counties in North Carolina; Cuyahoga and Lorain Counties in Ohio; and Allegheny County, Pennsylvania.
3. Researchers argue that ballot measures, too, should be written at the eighth-grade reading level. We did not include a Flesch-Kincaid analysis of ballot measures because their language is often not at the discretion of election administrators (Ballotpedia n.d.).
4. Microsoft Word (n.d.) and ReadabilityFormulas.com (n.d.) use the same formula for calculating reading levels. The formula for the Flesch-Kincaid grade level score used was (.39 x ASL) + (11.8 x ASW) -- 15.59, where: ASL = average sentence length (the number of words divided by the number of sentences) and ASW = average number of syllables per word (the number of syllables divided by the number of words). Discrepancies between the two scores can be explained by differences in their algorithms' calculation of the number of syllables and in how they handle punctuation, contractions, and so on (Neuhoff et al. 2016).
5. The ballot instructions of Lorain County include images of filled-in ovals but lack an illustration such as that shown in Figure 1.
6. Election administrators faced multiple challenges designing Maricopa County's 2024 general election ballot. In the one split contest on the ballot, there are 42 candidate names. Given the physical size of the ballot, the candidate names are split into two columns. The ballot is two front-and-back pages due to the number of issues and contests that needed to appear on the ballot (Fifield 2024). This may explain the design issues found in the instructions, as images that are likely of Maricopa County's 2018 ballot show more comprehensive and differently formatted instructions (Eltohamy 2020). Interestingly, the earlier ballot appears to also avoid center alignment and all capital letters for candidate names and offices (Eltohamy 2020).
7. While the option for straight-ticket voting (STV) is outside the discretion of election administrators, it has been found to decrease ballot roll-off (when voters fail to record choices on races/issues further down the ballot), decrease wait times to vote due to long ballots, and increase turnout (Engstrom and Roberts 2020; Spencer and Markovits 2010). Interestingly, several states have abolished STV since 2000, including two states analyzed in this report---North Carolina and Pennsylvania (Engstrom and Roberts 2020). Moreover, Michigan temporarily abolished the option for the 2018 general election before Michiganders voted to reinstate STV with a constitutional amendment (Engstrom and Roberts 2020).
8. Although Cuyahoga County's instructions are positioned on the left side of the ballot, the instructional text is not limited to the top corner.
9. On page 5, the options "yes" and "no" under Issue 1 are formatted in all capital letters.
10. Lorain County's ballot instructions are on the left side of the ballot; however, the instructional text spans beyond the top corner.
11. The sentence "Do not use red ink!," the phrase "Shall the amendment be approved?," and the options of "yes"/"no" appear in all capital letters.
12. Allegheny County's instructions are positioned on the left side of the ballot, but the text is not restricted to the top corner.
Related resources
Attacks on Science
Scientific Societies Call on Congress to Save Science
2024 Annual Report
