Young voters and minorities were more likely to have their ballots rejected by elections officials checking to see if the signature on the envelope matched the one on file, a new state report on the 2020 election concluded.
The total number of ballots rejected in Washington was small – slightly less than 30,000 out of more than 4.1 million cast – and the main factor in determining whether a ballot was rejected wasn’t age or race, but the county where a voter lived, the performance audit concluded.
Even then, auditors couldn’t point to a specific pattern that explained the differences in rejection rates.
Some were rejected because they were mailed after election day. Some came in envelopes with no signature, and some had signatures on the envelope that didn’t match the one on file.
Whether the report from the state auditor will satisfy anyone with questions about Washington’s all-mail voting system remains to be seen. It suggests ways to reduce the number of rejected ballots through better voter education and outreach to some communities that were statistically more likely to have their ballots rejected.
But it skips over a key reason for the rejections.
“A lot of those are ballots that should be rejected because they aren’t from the registered voter,” Spokane County Auditor Vicky Dalton said.
Spokane was not one of the 10 counties that were studied in the performance audit, although elections officials from all 39 counties were involved in discussions, Dalton said.
The signature on every ballot envelope is checked by people trained by the Washington State Patrol before the envelope is opened and the ballot removed to be counted. In some counties, several people compare the envelope signature to the signature on file at the same time. In Spokane County, one elections official checks the envelope and if he or she has doubts, it is checked by a second official with access to other signatures on file, and sometimes goes to a third official before deciding whether it’s a match.
Signatures can change over time. If the signature on the envelope is still in doubt, the voter is sent a letter asking them to submit a current signature that will be kept on file. If time is running out before the vote count is final, they’ll get a phone call to come to the office to sign a form.
The state average for all rejections was 0.72 of 1%. Auditors compared the five counties with the highest percentage of rejections and five with the lowest to look for trends. Spokane was below the midpoint, with a rejection rate of 0.46 of 1%.
But auditors couldn’t find any practices or procedures that the counties with the lowest rates had that the ones with the highest rates didn’t have, or vice versa. “The lack of one identifiable cause suggests that multiple factors affect the rate and no one practice is responsible,” auditors said.
They looked at signatures on a random sample of some 7,200 ballot envelopes and had hand-writing experts compare them to the signatures on file to see if they came up with a different conclusion than the county elections workers who decided whether to count or reject the ballot. The outside experts agreed with the decisions made by county workers nearly 99% of the time.
Ballots from voters aged 18 to 21 had a rejection rate of 2.68%, nearly four times higher than that state average. Those 22 to 25 had a rejection rate of about 2%, and the rates drop steadily after that so that those 65 and over had a rejection rate of 0.17 of 1%.
That’s not a surprise, Dalton said, because a person’s signature is often not set until their late 20s or early 30s. It’s something elections officials have experienced for years.
Men were more likely to have their ballots rejected than women, which was different from studies in two other states, auditors said. A study in Georgia showed women in that state were more likely to have their ballots rejected, while in Florida there was no significant difference between men and women.
While voter registration forms reveal age and gender, they don’t list race or ethnicity. Auditors used a formula that combines geography and surname information to predict the probability that the voter is a member of a racial or ethnic minority, and concluded minorities were more likely to have their ballots rejected than white voters.
White voters had an estimated rejection rate of 0.63 of 1%. The rejection rate for Asian Americans and Pacific Islanders was twice as large, for Hispanic and Native American voters about two and a half times larger and for Blacks nearly four times larger.
Auditors looked for any evidence of bias in the random sample, but noted that most signatures were either a clear match or easily rejected. They looked then at the signatures where the matches were “inconclusive” and found no patterns of bias that would lead to a higher rejection rate.
“Some racial and ethnic groups may face language or cultural barriers that increase the likelihood of ballot rejections,” auditors wrote. Previous studies suggested that people with limited ability to read English may make more mistakes on the ballot and have more trouble curing those mistakes, while signing a name in a new language may lead to inconsistent signatures.
Auditors suggested that elections officials contact voters in their preferred language when trying to “cure” signatures that don’t match. That’s not as simple as it sounds, Dalton said.
“We don’t have a preferred language,” she said. There’s nothing on the voter registration form to suggest that, and different immigrant communities have different traditions.
Among the Marshallese, she said, voters prefer to get the information from their elders rather than a translated form from the government.
“You need to work with each one of these communities to find out how they function and what they want,” Dalton said.