No, absentee rejection rate in November was not a “stunning anomaly”

The November election has come under attack on a variety of fronts from a variety of sources. Former President Donald Trump and his backers have taken aim at voting machines, vote reporting, vote counting, the mail-in voting process and more.

Many of those claims centered around Wisconsin, a key swing state that backed Trump in 2016 but supported Democrat Joe Biden in 2020. Election officials in the state have been unanimous that it was a free and fair election, with no evidence of fraud outside the usual scattered cases — and certainly no fraud that would have reversed Biden’s 20,000-vote margin of victory. Not to mention the recount that upheld the margin of victory.

Three months later, many remain unconvinced.

Conservative radio host Dan O’Donnell raised a new claim about election integrity in a Feb. 3, 2021, commentary piece written for the MacIver Institute, a conservative think tank based in Madison.

In a piece headlined, "Wisconsin’s Miraculous Vanishing Ballot Rejection Rate," O’Donnell asserted the low number of absentee ballots rejected in the November 2020 election was a "stunning anomaly."

He implied something questionable was at play.

"This is nearly impossible to explain," O’Donnell wrote. "Did hundreds of thousands of Wisconsinites suddenly figure out how to correctly fill out their ballots? Or was there a concerted effort to count as many ballots as possible without regard to whether they were correctly filled out? The evidence points toward the latter."

O’Donnell’s core claim has two elements: That the rate of absentee rejection was an anomaly, and that the tally should be considered questionable — or even evidence of a conspiracy of some kind.

We took a closer look at both elements.

O’Donnell’s claim runs into immediate trouble when we look at the historic trend of absentee ballot rejections.

His commentary included a chart showing six elections since April 2016, with four having an absentee rejection rate above 1.4%, and the November 2018 and November 2020 elections having a rejection rate of 0.2%. O’Donnell described this as evidence "ballots suddenly stopped getting rejected in the two statewide races in which unpopular Republicans (Scott Walker and Donald Trump) were on the ballot."

But zooming out farther shows the November 2020 numbers were to be expected since November elections have consistently lower rejection rates for absentee ballots.

From 2008 to 2020, clerks rejected an average of 0.62% of absentee ballots in November elections, according to Wisconsin Elections Commission data. The August, April and February elections ranged from 1.4% to 2.2%.

In addition, the rejection rate in November elections has steadily dropped over time, decreasing every November but one from 2008 to 2020.

The outlier election was in November 2016, due to a change in state law. In 2016, the state Legislature passed a law stating an absentee ballot may not be counted if it is missing the address of a witness. Every absentee ballot in Wisconsin must be signed by both the voter and a witness, who must also provide an address.

As a result, the absentee rejection rate jumped from 0.31% in November 2014 to a 1.35% in November 2016.

But election officials had become concerned some voters wouldn’t be aware of the witness address requirement, so in October 2016 the commission — on a motion made and seconded by Republican members — voted to advise clerks they should fix missing address components based on "reliable information." This guidance passed very close to the November 2016 election, so it wasn’t followed widely enough quickly enough to affect the rejection spike in that election, WEC spokesman Reid Magney said. 

Once clerks had time to incorporate the new guidance, the November rate resumed its descent — to 0.23% in November 2018 and 0.2% in November 2020.

In other words, limiting the timeframe to 2016-20 as O’Donnell did treats the November 2016 number as a baseline and the November 2018 and 2020 numbers as outliers. Putting that data in context makes it clear 2016 is actually the outlier.

And, of course, November 2020 is in line with similar elections in the past.

Featured Fact-check

Robin Vos

stated on November 6, 2024 in News conference

“Every single county in America, every single county, Kamala Harris did worse than Joe Biden did.”

That brings us to the question of why November elections — in 2020 and before — would have lower ballot rejection rates than other elections.

This is essentially the question O’Donnell’s raises throughout his column, but he never contacted state election officials to seek out the answer, Magney said. It turns out, it’s a pretty straightforward explanation.

The difference in ballot rejection rates is primarily because of how people vote. November elections bring a surge in early voting, which is still counted as absentee but takes place in a clerk’s office with the clerk as the witness.

"This eliminates chances that their ballot will be rejected for an insufficient (witness) certification," Claire Woodall-Vogg, executive director of the Milwaukee Election Commission, told PolitiFact Wisconsin in an email.

Looking at 2016 to 2020, for example, elections commission data shows in-person absentee votes accounted for an average of 19% of all votes in November elections, but only 7.3% of the vote in the April elections — even with the spike in in-person early voting in April 2020 due to the pandemic. (We also checked elections further back to confirm this was an ongoing trend.)

For those using traditional mail-in ballots, Woodall-Vogg said Milwaukee elections "cured" — that is, used reliable information to fill out missing witness information, per commission guidance — a total of 1,063 ballots in November. That generally meant simply filling in "Milwaukee, WI" for voters who included a street address but forgot the city and state and could be confirmed as living there.

"We wouldn’t make presumptions about a witness’s address; if we couldn’t read the witness’ name or there were multiple people by the same name, we would either call the voter or return their ballot via mail," Woodall-Vogg said.

Why did the rejection rate continue the downward trend in November 2020 despite the high volume of mail-in absentee ballots?

Magney noted some improvement in rejection rate would be expected from April to November 2020 since in April many voters were casting a mail-in ballot for the first time and in a short timeframe, while in November they had more time to find witnesses and return ballots, and clerks had more time to send deficient ballots back to voters to be corrected. 

Woodall-Vogg also noted clerks around the state were taking steps to educate voters given the expected influx of absentee voting.

"Election administrators and community groups all made an intentional effort to educate voters on the requirements when voting by mail due to the increase with COVID," Woodall-Vogg said in her email. "The WEC created new absentee instructions which were finally in plain English and even included photos. The media covered the requirements far more than ever before, as did we via outlets like Facebook and media interviews. On the exterior of our drop boxes, we had large stickers with STOP signs, asking voters to check if they signed, their witness signed, and their witness provided their address."

Finally, Magney noted the rejection figures cited by O’Donnell are only one way to handle flawed absentee ballots, making it an imperfect measure of how many absentee ballots ultimately don’t count. If a ballot is to be thrown out, some clerks mark this as a rejected ballot, while others record it an "administrative ballot cancellation." While the number of absentee ballot rejections dropped from April to November of 2020, the number of administrative ballot cancellations nearly quadrupled.

"The assignment of these labels is entirely at the discretion of local clerks," Magney said in an email.

Before we move on to the ruling, let’s pause for a moment to take a wider view.

This is far from the first conspiracy claim involving the 2020 election. And it’s a reminder that a good first step when examining any claim of coordinated wrongdoing is to consider the scope. If there were something nefarious at play, who would have to be in on it?

The answer in this case stretches well into the quadruple digits.

Wisconsin’s elections are overseen by the Wisconsin Elections Commission, but we’re not talking about a generic policy in this case. The rejection rate is the accumulation of thousands of individual, local decisions on whether (and how) to reject a given ballot.

Those decisions are made by the 1,850 municipal clerks around the state, and the thousands more deputy clerks and volunteers they work with.

Asked if he was alleging a coordinated effort across a group of nearly 2,000 clerks, O’Donnell said, "Clearly ballot mistakes that were accepted in these elections would not have been, and were pretty obviously not, accepted in other elections."

O’Donnell asserted in an opinion piece that November election data shows a "stunning anomaly in the rejection rate of absentee ballots."

But that rejection rate was neither stunning nor an anomaly.

Election data from 2008 to 2020 shows November elections have consistently lower rejection rates than other elections, due primarily to the higher number of in-person early voting (which is included in absentee tallies). And the November rejection rates themselves have steadily dropped in each election aside from 2016, when a new law caused a one-year spike. The November 2018 and 2020 rejections rates were in line with that long-term trend.

We rate this claim False.