In the days and weeks after the 2024 general election, the work did not stop for election officials across the country.
Many of them were required to do some form of post-election audit.
Audits come in many shapes and sizes, but all exist for the purpose of making sure that the results from the tabulation machines were correct.
Different Types of Audits
Across the country, our team observed automated audits, traditional or fixed percentage audits, and risk-limiting audits of the 2024 general election results.
Automated Audit in Lake County, Illinois
In Lake County, Illinois, elections staff conducted an automated audit by re-tabulating 5% of all ballots using a different scanner. No ballots were tallied by hand. Space was tight for observation, but our team member was present to view the automated audit.
Fixed Percentage Audit of Ballot Images in Honolulu, Hawaii
In Honolulu, Hawaii, state elections staff counted ballots from 10% of precincts in the state for their post-election audit held in the Senate Chambers of the Hawaii State Capitol. The space was wide open and accessible for observation, though very, very cold.
The audit could not be conducted until all ballots were incorporated into the final tally. This is why the audit took place on the last day to "cure" mail ballots: ADD DATE.
Instead of doing a manual tally using physical ballots, state election staff relied on digital images of ballots. Observing an audit using digital images of ballots was a first for our team. We joined a small number of observers, either selected by their political party or unaffiliated, who were on hand to view the process — and received a stipend for observing the audit. There was an observer "lead" available to provide instruction, if needed. The observers seemed to be familiar with the process because staff did not provide them with any pre-instruction. Observers could view tally sheets and had access to a "Counting Center" handbook that describes the counting process. The observers were talkative and nice, and the staff was very polite and focused on the task at hand.
Hawaii elections staff used a "calling off and tallying" method, which seemed to work well. The number of precinct ballots varied by the batch but there were typically five to 20 ballots in each batch. However, some batches only contained one or two ballots. For these tiny batches, staff spent more time on the keyboard transitioning to the next batch, which added to the time it took to audit these precincts.
Elections staff used five counting stations with typically two people at each station. One staff member would review the digital ballot and call off the responses or candidates for each race to be audited for that precinct. The other staff member would tally. From time to time, one staff member performed both functions.
The teams were not bipartisan and each team performed their functions a little differently. Some observers, including our colleague, wondered why one team was doing something differently from another team. For example, one team completed the portions of the tally form first by filling in the batch number and the number of ballots in the box. Other teams filled out the individual batch tally form and the cumulative report form all at once.
Efficient 5% Re-Tabulation Audit in Peoria, Illinois
The executive director of the Peoria County Election Commission led a post-election fixed percentage audit of 5% of ballots at the commission's main office. Our team member was the sole observer despite invitations to both partisan and nonpartisan observers and a long history of audit observation in Peoria.
With a small number of elections staff members participating, the audit was particularly efficient because the executive director had batched the ballots two days prior to the audit while conducting a dry run.
This was a machine re-tabulation audit that generated printed reports, but the commission did not prepare a public summary or report and did not have handouts describing the process for observers.
Our team member had a positive experience and appreciated the opportunity to think about the way Hart InterCivic tabulation equipment reads ballots. There was no need for a cast vote record because marks that did not signal requests for review or adjudication were read as "Yes" votes, not overvotes.
A First in Hughes County, South Dakota
Our team member was there to observe the first post-election audits in South Dakota. In Hughes County, the county auditor led the fixed percentage audit, which was conducted by two bipartisan teams of two. Members of media and officials from the secretary of state's office were on hand to observe. There was no direct communication between the county auditor and the observers.
The audit teams used tally sheets prescribed by law, but were not supplied with detailed instructions for how to perform the audit. Instead, they were given suggestions on methods for reviewing and recording votes on the ballots they were auditing. It was up to each team to decide how to proceed. Our team would have liked to see written procedures or a quick reference guide for audit teams that observers could use to follow along.
The audit teams encountered an undetected reconciliation issue, which reinforced our team's desire to incorporate ballot accounting and reconciliation into the audit process. We were impressed by the county auditor's explanation of the importance of seals on ballot containers and how they work, and how the seal numbers are recorded and verified as part of the chain of custody.
Online Observation of the District of Columbia's Fixed Percentage Audit
Our team experienced its first Zoom observation opportunity when viewing the District of Columbia Board of Election's post-election audit on Nov. 22 and 25. There were approximately 12 observers online The ease of access was impressive, but the muted sound made for a less than full observation experience.
Eight audit teams showed no hesitation about the process. Each team had four members: two talliers, one caller and one witness to confirm counts. Teams were seated at tables in two large rooms, and the Zoom camera angle gave online observers a good view of each room. Staff brought the selected ballots to each table and provided each team with instructions, tally sheets and a list of each person's responsibilities. With the microphone active, there was a brief introduction before counting began. Staff encouraged the teams to take their time and count accurately. Staff instructed the teams to alert them if voter intent was unclear so staff could determine how the a ballot was marked. Staff was highly engaged and on their feet, checking in with all tables throughout the audit.
Highly Organized 1% Manual Audit in Nevada County, California
In Nevada County, our team member had a positive and educational audit observation experience, buoyed by the elections staff’s professionalism and attention to keeping materials organized and chain of custody. With up to 15 staff members on hand to sort batches, select random numbers and pull batches, the audit started on time in the clerk’s office and two conference rooms in the county’s main building. The office had terrific signage throughout, making it easy to recognize areas of the office. Elections officials and observers had different colored badges, which highlighted the office’s commitment to organization and planning. The office used different colored bins for each stage of ballot processing, which, thanks to well-placed glass, was easy to view from public corridors.
While bipartisanship was not apparent to observers, two teams of four conducted the audit. Two team members reviewed voter selections and two others recorded the selections on tally sheets. Each team had one senior election official to oversee the process. All team members wore county identification.
Observers could view every phase of the audit, from the random selection of ballots through final tallies of voter selections. Three individuals observed the audit for an hour or more; our team member was the only one to observe the random selection and pulling of selected ballots. While there were no handouts for observers, there was a sign explaining a 1% manual tally.
During the audit, one team discovered a discrepancy between the batch report and the manual tally of the ballots in the batch. This triggered an investigation and a team discussion that was inaudible to observers. Our team noted that observers less familiar with post-election auditing and election administration would likely have had questions and perhaps felt skeptical about the discovery and investigation unless the county supplied additional information.
Careful Hand-Counting in Linn County, Oregon
Linn County’s post-election audit took place in the processing room in the elections office in Albany, Oregon. The county clerk, chief deputy clerk, records supervisor, elections supervisor, elections staff and temporary elections staff conducted the audit. The room was large, but long and narrow, so space was limited. The observation area was at the front of the room, and there was a TV screen displaying camera feeds from the processing room. With tables set up along the length of the room, observers could really only see the first table. Each table had five people: four bipartisan workers and one supervisor.
The staff was polite and kind, and the county clerk was engaging and ready to answer questions. Workers wrote their counts on notepads instead of pre-printed audit materials such as tally sheets, and supervisors used printed batch results to mark off contests as they were counted. Teams had green papers to signal when they needed assistance. Our team member observed how Oregon’s audit processes differ from other states' and learned about how statutory requirements compare with best practices.
The hand-counting pace was intentionally slow to avoid mistakes. Overall, the audit was uneventful and successful. The teams found no discrepancies between the election results and their hand-counts.
Confidence and Bipartisanship in Clackamas County, Oregon
In Clackamas County, Oregon, the county clerk, elections director, elections staff, temporary staff and party volunteers audited three races in the election office’s processing space. There were six tables, each with pairs of workers affiliated with different political parties. Workers wore colored lanyards to identify their party affiliation or if they were unaffiliated.
Our team member was one of three observers, viewing the audit through large windows lining the hallway. Just two tables were fully observable given the room’s layout. The elections director frequently visited the observation area to answer questions and discuss the elections process with observers. The staff was very kind and confident in the overall process and conducting the audit. There did not appear to be any confusion or clarification needed as to their roles and responsibilities.
For each race being audited, workers sorted their batch of ballots into piles for each candidate or issue choice, overvotes, undervotes and adjudicated ballots. Then they counted each and wrote their totals on sticky notes. When they finished, workers reported their totals to a supervisor who used printed batch summaries to compare results and confirm whether the hand-counts were accurate or not. If not, they repeated the process. Our team member did not observe any inaccurate counts and noted that the hand-counting was slow with processes built-in to prevent mistakes.
Three Types of Audits in Bartow County, Georgia
Led by the supervisor of elections, the Bartow County Board of Elections conducted a risk-limiting audit and a hand-count audit of the presidential contest. They also worked with Enhanced Voting to pilot an automated audit using ballot images.
The audits took place at the senior center and the county encouraged bipartisan observation. Observers included our team member; colleagues from Verified Voting, Voting Works and Enhanced Voting; staff from the secretary of state’s office; and political party observers.
There was ample space for observation, but no public-facing explanations of the process or handouts to help observers follow along. Contentious observers were present for Bartow County’s 2020 and 2022 post-election audits. In 2024, observers were amicable and the observation experience was positive.
Innovations in Audits
In Georgia, election officials worked with Enhanced Voting to conduct a ballot image audit.
In Honolulu, Hawaii, all of the staff used earpieces for their walkie-talkies. This was both useful and noteworthy from a security perspective.
The District of Columbia Board or Elections provided a Zoom link for public observation of its post-election tabulation audit. This approach increased access to the process but with the sound muted, the observation experience was limited. Our team hopes the Board will unmute for future audits for a full observation experience. If that proves impractical, periodic updates and process explanations from staff — or an observer guide — would greatly enhance the online observation experience.
Looking Ahead
As officials look beyond the 2024 election cycle and toward the future, our team is taking a deeper dive into audits. We want to help uncover and address common issues and increase awareness of the types of audits and processes that are working well for different offices.