Close Close Comment Creative Commons Donate Email Add Email Facebook Instagram Mastodon Facebook Messenger Mobile Nav Menu Podcast Print RSS Search Secure Twitter WhatsApp YouTube

Handwriting Disputes Cause Headaches for Some Absentee Voters

Election board procedures for verifying signatures vary widely. Lawsuits in several states argue this deprives people of the right to vote.

An election clerk in Sacramento inspects a mail-in ballot earlier this year. (Rich Pedroncelli/AP Photo)

Elisabeth Warner drove 12 hours the weekend before Election Day so her daughter could vote.

Her daughter, Emma Warner-Mesnard, 20, attends Transylvania University in Lexington, Kentucky, but is registered to vote in her hometown, Columbus, Ohio — about a three-hour drive away. Warner-Mesnard mailed the Franklin County Board of Elections her request for an absentee ballot in mid-October, but officials there determined her signature on the application didn’t match what they had on file and placed her ballot on hold.

Warner-Mesnard sent in a second absentee ballot application, which the county accepted Oct. 31. Aaron Sellers, a public information officer for the Board of Elections, said her ballot was on track to be delivered Saturday, giving Warner-Mesnard just enough time to get it in the mail by Ohio’s Monday postmark deadline. But the campus post office isn’t open on weekends, and she wasn’t willing to take any chances. So the mission to vote became a road trip: On Saturday, Warner-Mesnard’s mother picked her up in Lexington and drove her to Columbus to cast an early ballot in person at the Franklin County Board of Elections. On Sunday, her mother ferried her back to Lexington, spending a total of about 12 hours on the road.

“It’s just annoying that that’s the way that it has to go,” Warner-Mesnard told ProPublica.

She’s one of tens of thousands of voters facing issues over their handwriting, according to data from the U.S. Election Assistance Commission. According to a report by the Election Assistance Commision, 99 percent of the 33 million absentee votes submitted for the 2016 election were ultimately counted, but of those rejected, a “non-matching signature” was the No. 1 reason, beating out tardy ballots.

Ohio, like many states, requires election officials to compare a voter’s signature with existing records before issuing and counting absentee ballots. But voting rights advocates say that a lack of sufficient standards around signature verification can disenfranchise people, and that the rules disproportionately affect voters who aren’t native English speakers. Two lawsuits filed by advocacy groups in Georgia, where allegations of voter suppression have dominated the hotly contested governor’s race, argue that election officials aren’t adequately trained to verify signatures and don’t give voters proper notice to fix handwriting issues.

One of the suits was brought by the Lawyers’ Committee for Civil Rights Under Law, which provides data to Electionland.

The groups are suing Secretary of State Brian Kemp, who is the Republican candidate for governor, and Gwinnett County, which had rejected close to 1 in 10 of the mail ballots it received as of Oct. 12, shortly before the suits were filed. This amounted to around 37 percent of the ballots rejected in the entire state. Not all were nixed because of alleged signature mismatches. Some voters forgot to sign or appropriately date their ballots, or left off other required information. Some blame such errors on the design of the county’s return envelope, issued in English and Spanish for the first time this year.

An Oct. 25 temporary restraining order applied to both lawsuits required Georgia counties to notify voters before rejecting absentee ballots for alleged signature mismatches and to process them as provisional ballots. This year, the ACLU won cases against New Hampshire and California over their signature verification rules. The Florida Democratic Party and the Democratic National Committee prevailed in a similar suit in Florida in 2016.

Verifying a voter’s signature on a mailed ballot is a key part of election security, similar to the way you sign a poll book when you vote in person, said Amber McReynolds, the executive director of the National Vote at Home Institute. However, she says standards around signature verification vary widely by state.

Colorado tightened its processes after forged signatures on petitions came to light in 2016, leading the state to implement a system McReynolds called “the gold standard.” Officials there have another strong incentive to get things right: The state conducts all its elections via mail.

In Denver, where McReynolds was formerly head of elections, officials first run mail ballots through signature verification software that automatically matches around 30 to 45 percent of the signatures with existing records for the voter. A team of bipartisan election judges with basic training in signature verification look at the remainder. If they suspect someone’s handwriting doesn’t match previous samples, the case gets flagged for a second team to review. If the human reviewers agree that the signature doesn’t seem to match, they immediately send the voter a text or email in addition to a mailed notice, so they can quickly fix the problem by providing valid ID. In the last election, McReynolds said only 0.8 percent of Denver voters had a signature discrepancy that wasn’t resolved.

“The software is agnostic to anything about the voter, which is why it’s great,” McReynolds told ProPublica. She says this makes the process more consistent, because “people see things differently.”

However, in many jurisdictions, election officials with little or no training in verifying a person’s signature are tasked with doing just that.

“That’s scary,” said Mark Songer, an FBI-trained handwriting analyst and forensic document examiner who has trained Colorado election workers in signature verification. He told ProPublica that handwriting changes over time and that no two signatures from the same person are ever identical. Health, eyesight, medications and the writing surface, paper and pen used can affect how a signature looks. To determine a match, it’s important to compare several instances of someone’s handwriting; it’s unlikely that only one or two samples will show the spectrum of a person’s normal variations.

Moreover, distinguishing signatures isn’t an exact science. “Even major treatises on handwriting analysis concede that it is extremely difficult for anyone to be able to figure out if a signature or other very limited writing sample has been forged,” Jonathan Koehler, a professor and forensic science expert at Northwestern Pritzker School of Law, told FiveThirtyEight last year.

In Franklin County, Sellers said election records showed Warner-Mesnard had a few different signature styles in the last few years, and the one she submitted with her first application didn’t resemble what was on file. “I can certainly understand why that would have been put on hold,” he said, noting the Board of Elections tends to give voters the benefit of the doubt. “To decline a signature, that’s not something that is taken lightly.”

Warner-Mesnard said it was possible her signature had changed, but the process to rectify the issue frustrated her. Sellers said the Board of Elections received Warner-Mesnard’s initial absentee ballot application on Oct. 22 and promptly mailed a letter to her college address notifying her of the signature mismatch. She says she never got it and only learned of the flag when she and her mother called the Board of Elections asking why her ballot hadn’t arrived.

A passionate voting rights activist, Warner-Mesnard was willing and able to go the literal distance to make sure she could cast a ballot. But many who plan to vote absentee cannot come in person because of disability, illness, travel, work or family obligations. Some voters flagged for signature issues may be deterred by the additional steps required to “cure” a mismatch.

Maxine Witteveen, an 18-year-old freshman at Georgia Southern University, was excited to vote for the first time. She and her mother, Laila, were surprised when a letter arrived from the Cobb County Board of Elections stating that the signature on Maxine’s absentee ballot envelope “does not appear to be that of the voter.” Laila says the county told her that election officials compared her daughter’s cursive handwriting on the oath with the one on her driver’s license, where she printed her name. Songer, the handwriting expert, agrees this was a clear reason to flag the ballot.

Laila says the county initially advised Maxine to re-apply for an absentee ballot and try to replicate her handwriting as it appeared on her license. But after the court order, Marilyn Marks, executive director of the Coalition for Good Governance, which underwrote one of the lawsuits, contacted the family and said Maxine could provide a copy of her ID to the Board of Elections, and her absentee ballot would count. Election officials emailed an affidavit for her to sign and return with a scanned copy of her ID.

Laila is grateful for the county’s assistance — and Marks’ nudge. She’d helped her daughters navigate the absentee paperwork and stamped their envelopes because she really wanted both to vote. But without that reminder to fix the signature issue, she says, “We might have let it go.”

About Electionland

ProPublica’s Electionland project covers problems that prevent eligible voters from casting their ballots during the 2020 elections. Our coalition of newsrooms around the country are investigating issues related to voter registration, pandemic-related changes to voting, the shift to vote-by-mail, cybersecurity, voter education, misinformation, and more.

Questions? Read our FAQ.

Follow Electionland


and 150+ local and national newsrooms. Sign up to become a partner here.

Technical Partners

More Election Tools

The User’s Guide to Democracy

Congress works for you. Here’s how to be a better boss.


See what your representatives in Congress say and do.

ProPublica on IFTTT

Do more with ProPublica data and automated notifications.

Latest Stories from ProPublica

Current site Current page