In-Person Voting in 2024
How Voters Interact with Administrative Processes on Election Day
Even though mail ballots have become an increasingly popular way to vote, many individuals still cast their ballots at their local polling place or vote center. When doing so, these voters interact with a different set of administrative processes managed by local election officials than those who vote by mail.
As the 2026 midterm election draws closer, we take a look back at what the last Elections Performance Index (EPI) tells us about the in-person voting experience in 2024, including accessibility, voting wait time, and provisional ballot usage.
Accessibility Barriers
How this Indicator is Measured
In a study of the 2016 election, the Government Accountability Office found that 83% of polling places had one or more potential barriers to voting, and nearly two-thirds had an accessible voting station with at least one feature that could keep a voter with a disability from casting a private and independent ballot. In part, this illustrates the challenges election officials face in trying to find locations that meet the minimum requirements to be a polling place, and why so many polling places are schools or public auditoriums.
The disability access indicator “measures the degree to which voters are deterred from turning out to vote because of disability or illness.” Notably, this indicator is the only measure in the EPI calculated using percentage points. After an update to the original EPI indicator in 2020, the disability access indicator uses the Voting and Registration Supplement (VRS) to the Current Population Survey (CPS) to measure the difference in turnout rates between people who reported having one of six disabilities listed in the CPS and those who reported having none of these disabilities.
By the Numbers
In 2024, the average nationwide score was -7 pp away from zero, meaning that there was a 7 percentage point gap in turnout between eligible voters who reported to the VRS they had one of six disabilities and those who did not. Had turnout been equal, roughly 980,000 more people with disabilities would have voted in 2024. The nationwide average score for this indicator improved by two percentage points from 2020, when turnout was likely impacted by the COVID-19 pandemic and the changes to in-person voting that necessitated—directly affecting how voters experienced polling place accessibility.
The figure above compares changes in the disability access indicator by state between the 2020 general election and the 2024 general election. Dark green indicates that the turnout gap between voters with disabilities and voters without was smaller in the 2024 general election than the 2020 general election. Dark gold indicates that the turnout gap between voters with disabilities and voters without was larger in the 2024 general election than the 2020 general election. Gray indicates that there was little to no change in the turnout gap in 2024 compared to 2020.
The figure above compares changes in the disability access indicator by state between the 2020 general election and the 2024 general election. Dark green indicates that the turnout gap between voters with disabilities and voters without was smaller in the 2024 general election than the 2020 general election. Dark gold indicates that the turnout gap between voters with disabilities and voters without was larger in the 2024 general election than the 2020 general election. Gray indicates that there was little to no change in the turnout gap in 2024 compared to 2020.
There is a measurable turnout gap between voters who self report having a disability and those who decline to self report: an Election Assistance Commission and Rutgers report from 2022 found that 20% of those who self-reported as having a disability expressed difficulties with voting in person, compared to 6% of those who said that they experienced difficulty voting in person, but did not self-report as having a disability. These reported voting difficulties can be caused by a wide variety of reasons, including poor signage, challenging layouts, long lines, or the lack of a private place to cast a ballot. Coupled with potential issues voters face when registering to vote, as well as other factors such as socioeconomic status, education, access to transportation, and age, individuals who self-report as having a disability can face difficulties at a variety of touchpoints when going to cast a ballot.
Voting Wait Time
Election officials and researchers have spent countless hours studying how best to deploy election staff and resources to minimize the amount of time any one voter must spend waiting to vote. News reports and voters often complain about long lines on Election Day; indeed, these lines can be a potential barrier for voters who wish to cast their ballots in person or return a mail ballot to their polling place. Long lines can indicate excitement about a particular election, but variations outside of the norm can suggest other administrative factors are at play (for example, insufficient signage, voter roll issues, or an increase in the number of provisional ballots). The average wait time is one quantifiable way to measure the ease of voting: shorter wait times may indicate a more convenient, accessible experience for voters.
How this Indicator is Collected
This indicator is calculated using data from the Survey of the Performance of American Elections (SPAE), which is a national survey of registered voters focused on understanding the voting process from the voter’s perspective. The SPAE collects information from voters in each state about their Election Day and voting activities, giving us a clearer picture of voters’ experience in each state, including how long they waited to vote.
While in-person voters are the most obviously affected by long wait times, this indicator also includes data from mail voters (including voters in all-mail-voting states). While that may seem counterintuitive at first, mail voters who drop off their completed mail ballots in-person at a polling place, for example, may also report waiting in a line to cast their ballot.
How Long it Took to Vote In-Person
A 2014 report from the Presidential Commission on Election Administration, spurred by reports of long lines in the 2012 election, laid out a recommendation that no voter should wait more than 30 minutes to cast a ballot. In 2024, the nationwide average wait time to cast a ballot was 10.4 minutes. States with the shortest wait times included Oregon, Vermont, and Rhode Island, while Indiana counted the highest wait time for voters.
This figure depicts the average voting wait times in all 50 states and the District of Columbia, as captured by the Survey of the Performance of American Elections, in all presidential elections since 2016. States are sorted according to their average reported voting wait time in the 2024 general election; states appearing at the top of the figure had the highest average voting wait time in the 2024 general election, while those appearing at the bottom had the lowest.
This figure depicts the average voting wait times in all 50 states and the District of Columbia, as captured by the Survey of the Performance of American Elections, in all presidential elections since 2016. States are sorted according to their average reported voting wait time in the 2024 general election; states appearing at the top of the figure had the highest average voting wait time in the 2024 general election, while those appearing at the bottom had the lowest.
Long wait times at polling places have come under intense media scrutiny in recent years. But, because wait times can be caused by a variety of reasons, it is important to evaluate wait times within the unique electoral landscape of each state alongside the individual voter experiences examined in the SPAE.
Provisional Ballots
What are Provisional Ballots?
During the 2000 election, there were reports of eligible voters who were turned away at their local polling place and were unable to cast a ballot. At the time, only 17 states and Washington D.C. had some kind of provisional ballot requirement; six additional states offered same-day voter registration options. When HAVA was passed into law in 2002, it required states to provide an option for voters to use a provisional ballot, with exceptions for those states which allowed same-day voter registration. (Provisional ballots are ballots used by election officials when an individual goes to a polling place to cast a ballot but, for a variety of reasons, has an issue with their voter registration and otherwise would not be able to cast a ballot.)
Under HAVA, each state determines when a provisional ballot may be used, including in cases where there are discrepancies between voter registration records and what information a voter presents at the polls, if a voter goes to the wrong polling place, missing identification requirements, address changes, or voting in-person after requesting a mail ballot, just to name a few. The 2024 Election Administration and Voting Survey (EAVS) reported that 1,736,209 provisional ballots were issued in the 2024 federal election; approximately 1,281,399 were counted, at least in part, and approximately 436,258 were rejected.
Provisional Ballots Cast
The provisional ballots cast indicator tracks the number of provisional ballots used in a state, measured as a percentage of all votes cast. The usage rate is determined by dividing the total number of provisional ballots cast by the total number of votes. In 2024, the nationwide average for provisional ballots cast was 1.03%, about the same as the average of 1.04% in 2020. Overall use of provisional ballots in 2024 remained low, indicating that there are a small number of high-usage states pulling up the mean number of provisional ballots; most states have had a provisional rate of under 1.5% since 2008, with a few outliers, which are illustrated in the chart below.
The figure above shows states with a provisional ballot rate above 1.5% of total ballots cast in general elections since 2008. Dark green indicates a higher provisional ballot rate; light green indicates a lower provisional ballot rate. Gray indicates that the state reported too much missing data to calculate the provisional ballot rate for that election. Overall, the graph shows the small number of high-usage states that pull up the mean provisional ballot rate. We can see that the highest spike in the rate of provisional ballot use occurred in DC in 2012, which was the first year the District allowed Election Day voting. States not pictured exhibited a provisional ballot rate below 1.5% in all general elections between 2008 and 2024, or do not use provisional ballots.
The figure above shows states with a provisional ballot rate above 1.5% of total ballots cast in general elections since 2008. Dark green indicates a higher provisional ballot rate; light green indicates a lower provisional ballot rate. Gray indicates that the state reported too much missing data to calculate the provisional ballot rate for that election. Overall, the graph shows the small number of high-usage states that pull up the mean provisional ballot rate. We can see that the highest spike in the rate of provisional ballot use occurred in DC in 2012, which was the first year the District allowed Election Day voting. States not pictured exhibited a provisional ballot rate below 1.5% in all general elections between 2008 and 2024, or do not use provisional ballots.
In some cases, counties report cycles of provisional ballot usage: reporting none during one election, and considerably greater usage the next cycle. This can be due to a number of factors, from the size of the county itself to changes in election administration or overall voter turnout.
The figure above shows the nationwide average provisional ballot rate in all general elections since 2008. States that reported too much missing data to calculate the provisional ballot rate for that election were excluded from this analysis. For a list of these states, see the 2024 EPI Methodology Report. This graph also excludes five states—Idaho, Minnesota, New Hampshire, North Dakota, and Vermont—that do not use provisional ballots. Overall, the provisional ballot rate increased in 2024 compared to 2022, but remains on par with provisional ballot rates reported between 2014 and 2020.
The figure above shows the nationwide average provisional ballot rate in all general elections since 2008. States that reported too much missing data to calculate the provisional ballot rate for that election were excluded from this analysis. For a list of these states, see the 2024 EPI Methodology Report. This graph also excludes five states—Idaho, Minnesota, New Hampshire, North Dakota, and Vermont—that do not use provisional ballots. Overall, the provisional ballot rate increased in 2024 compared to 2022, but remains on par with provisional ballot rates reported between 2014 and 2020.
Provisional Ballots Rejected
The number of provisional ballots rejected represents the number of voters who cast a provisional ballot that was ultimately rejected for a variety of reasons. Each state has its own criteria for accepting or rejecting a provisional ballot, with one of the most common reasons for rejection being that a voter went to the wrong polling place.
The provisional ballot rejection rate is determined by dividing the number of rejected provisional ballots cast by the total number of votes. A majority of states that issue large numbers of provisional ballots verify those votes through curing processes after the election, and count them once they are cured. There is also a cyclical nature at the county level of provisional ballot usage, wherein one county might issue no provisional ballots one cycle, and many the next.
All in all, the rate of rejected provisional ballots has remained quite low and steady between 2016 and 2024, with a national average of 0.25% in 2024.
The figure above shows the nationwide average provisional ballot rejection rate in all general elections since 2008. States that reported too much missing data to calculate the provisional ballot rejection rate for that election were excluded from this analysis. For a list of these states, see the 2024 EPI Methodology Report. This graph also excludes five states—Idaho, Minnesota, New Hampshire, North Dakota, and Vermont—that do not use provisional ballots. Overall, the nationwide average provisional ballot rejection rate in 2024 was .12 pp above that of 2022, but only .04 pp above that of 2020.
The figure above shows the nationwide average provisional ballot rejection rate in all general elections since 2008. States that reported too much missing data to calculate the provisional ballot rejection rate for that election were excluded from this analysis. For a list of these states, see the 2024 EPI Methodology Report. This graph also excludes five states—Idaho, Minnesota, New Hampshire, North Dakota, and Vermont—that do not use provisional ballots. Overall, the nationwide average provisional ballot rejection rate in 2024 was .12 pp above that of 2022, but only .04 pp above that of 2020.
Impacts of Voting Provisionally
Because states can determine their own standards for administering and counting provisional ballots, their impact on election administration varies. On one hand, provisional ballots enfranchise eligible voters who may not have been able to cast a ballot on Election Day otherwise. On the other hand, provisional ballots must be individually verified or cured, adding delays to the counting process and leaving them subject to human, everyday clerical errors. Election officials can face additional scrutiny when processing them, or criticism that increased provisional ballot usage indicates a deeper problem with voter list accuracy and maintenance processes. Following Election Day, provisional ballots have often been the subject of intense litigation. (Our explainer on the subject dives deeper into research on provisional ballot usage in the United States.)
Voting In-Person in 2026 and Beyond
As Election Day 2026 gets closer and closer, voters are going to interact with the network of policies around accessibility, voting wait times, and provisional ballot usage in order to cast a ballot, especially if they vote in person. The next iteration of the EPI will continue to provide reliable, quantifiable metrics of the voter experience at these three touch points in the Election Day process, providing researchers with essential information measuring election administration nationwide.
