This is an online version of the executive summary. A PDF of the executive summary can be downloaded [here](https://www.ucs.org/sites/default/files/2026-03/Blocked from the Ballot Box_summary_2.27.26_0.pdf). A PDF of the full report can be found [here](https://www.ucs.org/sites/default/files/2026-03/Blocked from the Ballot Box_report_2.27.26_0.pdf). And an interactive map illustrating the report can be found here.
Structural Obstacles Depress Turnout, Exacerbate Ballot Rejections Across Racial Lines
Existing class and race inequalities in the United States mean that the foundations of political participation—resources, interest, and recruitment—are not equally distributed across groups. As a result, members of marginalized communities are often unable to participate effectively. Unfortunately, the structure of our political institutions and electoral system, as well as some electoral rules, can worsen these societal inequalities.
To assess the pervasiveness of racial inequalities, we examined turnout rates in the 2016, 2020, and 2024 general elections across precincts of varying race majorities in 11 battleground counties across seven states: Allegheny (Pittsburgh) and Philadelphia Counties in Pennsylvania; Columbus, Durham, and Mecklenburg Counties in North Carolina; Cuyahoga (Cleveland) and Lorain Counties in Ohio; Fulton County (Atlanta) in Georgia; Maricopa County (Phoenix) in Arizona; Milwaukee County in Wisconsin; and Wayne County (Detroit) in Michigan.
While our analysis finds both registered voter turnout and census voting age population (CVAP) turnout generally lower in precincts with a majority racial group of color than in majority-White precincts in every election year, examining turnout alone cannot fully capture the relative equality of elections and the inequalities in electoral power. Each election, hundreds of thousands of ballots are rejected. Common reasons for ballot rejections include missing or nonmatching signatures and late delivery. Ballot rejections are a natural and necessary component of elections, but research finds that ballots cast by members of certain groups, such as inexperienced voters, younger voters, and voters of color, are more likely to be rejected.
To facilitate analysis, we divided precincts into thirds—lower (lower number of rejections), middle, and upper (higher number of rejections). Our analysis finds inequities in that ballot rejection rates were higher in precincts with a majority racial or ethnic group of color than in majority-White precincts in 2016 and 2024. Interestingly, in 2020, there was less racial inequality across the three rejection categories, but majority-White and majority–Asian American/Pacific Islander precincts were proportionally more represented among the precincts with lower rejections.
Precincts with the lowest rates of turnout were also most likely to be in the high-incidence category of ballot rejections for all three election years. That is, communities with lower rates of turnout also experience higher rates of ballot rejection. As a result, these communities' political representation is diminished relative to those with high turnout and lower rates of rejection.
From 2016 to 2024, the percentage of majority-White precincts in the high-incidence category of ballot rejections remained fairly stable—around 20 percent. In 2020, however, this number was around 34 percent. The percentage of majority-Black precincts in the high-incidence category was about 30 percent in 2016, increased to about 37 percent in 2020, and decreased back to about 30 percent in 2024. The percentage of majority-Hispanic precincts in the high-incidence category decreased slightly from 2016 to 2024—from about 78 percent in 2016, to below 25 percent in 2020, and finally, to nearly 71 percent in 2024.
In 2020 and 2024, every other racial majority group was shown to have higher rejection rates than majority-White precincts. These racial disparities in voter turnout and ballot rejections—symptoms of systemic inequities in our electoral processes and political institutions—culminate in the underrepresentation of the interests, needs, and preferences of communities in majority-Black, -Hispanic, –Native American, –Asian American/Pacific Islander, and racially plural precincts as well as the overrepresentation of the interests of communities in majority-White precincts. Therefore, the decisions of elected officials and subsequent public policies are less likely to protect the health, safety, and well-being of these underrepresented communities.
Fortunately, there are ways to reduce current racial inequalities in elections and work toward a multiparty, multiracial democracy. More equitable ballot design, including redesigning mail ballots to be more user-friendly and expanding language and disability access, can increase turnout and lower rejections among marginalized groups. Increasing election data transparency, including adopting new voter file maintenance practices and expanding curing opportunities and outreach, can achieve similar outcomes. Currently, we are actively advocating for these and other science-based policies in target states.
The process of obtaining data for this analysis highlights the widespread absence of comprehensive and accessible precinct-level election data. Data were available in non-machine-readable formats, machine-readable formats without explanations on coding or data descriptions, or not at all. Moreover, it was often unclear which level of administrators—state or local—should be contacted to obtain these data. Both the availability and quality of election data are indispensable to the study of elections. Without access to these data, researchers are constrained in their ability to identify issues within the administration of elections or to develop solutions to overcome these challenges.
While our analysis is rigorous, our research and the resulting report—as well as the research of others in the field of election science and administration—would be improved with more accessible and usable data. This unavailability illustrates the continuing importance of the Union of Concerned Scientists’ work in the area of election data transparency, such as our development of election science recommendations for improving current levels of transparency.
During post-publication review, we identified some issues affecting the specificity of turnout and rejection calculations, which we have addressed. Although none of these changes altered the core findings of the report, they do refine our calculations and improve the overall precision of the analysis. In addition to these specific changes, the revised report contains a modestly expanded discussion of the challenges of this type of data collection, formatting, and processing. The errata can be found at www.ucs.org/resources/blocked-ballot-box.