The Washington PostDemocracy Dies in Darkness

One chain of fraud falsehoods extends back to a GOP Senate candidate

Kathy Barnette, Republican U.S. Senate candidate, speaks during an interview after the Pennsylvania Senate GOP primary debate in Harrisburg, Pa., on April 25. (Hannah Beier/Reuters)
Placeholder while article actions load

There was no reason to think Kathy Barnette would win her 2020 race for the U.S. House.

Barnette, a Republican, was facing incumbent Rep. Madeleine Dean (D) in a redrawn district near Philadelphia that Hillary Clinton had won by 20 points in 2016. Even if the race had occurred in a year when Republicans were heavily favored, a victory for Barnette would have been a stretch, and in 2020 in Pennsylvania, it was not a particularly good year for Republicans. Barnette lost and that, it would seem, was that.

But it wasn’t. Even before vote-counting was completed in the state, President Donald Trump began challenging the results. He insisted that his election-night lead was somehow the real tally and that the absentee ballots being slowly counted represented some scheme to upend his success. It was obviously false; it had been apparent for weeks that absentee ballots would favor Joe Biden and be counted after polls closed. The question was solely whether Trump’s lead (the “red mirage,” it was dubbed) would withstand the eventual count. Such clarifications were not useful to Trump, of course, so he ignored them.

So did Barnette. In an appearance on Fox News, she insisted she had won the day-of balloting but had to watch that be upended by “ridiculous” vote-counting after the fact. The process, she claimed, was akin to how elections were conducted in Afghanistan.

Barnette did not give up on elections. She is now seeking the Republican nomination for the Senate in Pennsylvania, hoping to appeal to the hard-right base of support that gave Trump almost enough votes to win the state. And to make that pitch, she can point to one particular contribution to the cause: the elevation of an election-fraud theory that has since metastasized into its own vibrant universe of surreality.

Sign up for How To Read This Chart, a weekly data newsletter from Philip Bump

It began not with the election but with the pandemic.

In late August, Barnette hosted a man named Douglas Frank for a discussion of the coronavirus in a video interview. Frank, an administrator and teacher at a school in Ohio, had compiled data assessing the progress of the pandemic over the course of 2020 and arrived at a determination that would certainly have pleased the president: The virus had run its course. A Facebook post making that point went viral, prompting the Associated Press to debunk it. (History has since debunked it further.) But Frank was making a pitch about reopening schools that Barnette found compelling. His case might have been flawed, but he made up for it with lots of visually interesting, complicated graphs.

This apparent expertise seems to have lingered with Barnette. When she decided after the 2020 election that something suspect must have happened, she contacted Frank. Frank’s analysis showed what Barnette might have hoped: There were tens of thousands of votes cast in Montgomery County that weren’t connected to real people. As the Philadelphia Inquirer reported last year, this appeared to confirm the other analysis Barnette had received. The fraud, she apparently believed, was real.

The problem, of course, is that it wasn’t. We’ve assessed Frank’s theory of the case before, including when it was later championed in Michigan by an attorney named Matt DePerno — now the presumptive Republican nominee for attorney general in that state.

In short, Frank took data on turnout by age from individual counties and then averaged rates across a handful of counties. He then compared that average to counties across the state, finding that they all correlated strongly — something he attributed to an unnatural introduction of an algorithm. But of course, there’s a much simpler explanation. Turnout patterns by age are consistent both within and across elections, with younger voters turning out less than older ones. Frank was simply representing that pattern. What he was doing was the equivalent of averaging four of Usain Bolt’s 100-meter dash times and announcing that the average was remarkably close to the times in his other races. Which: Of course.

But it all sounded complicated. He talked about “sixth-order polynomials” and had lots of charts and discussed r-correlation values. To a layperson, it seemed like an expert trying to convey complicated insights. To experts, it seemed like a layperson trying to feign complicated insights.

“I don’t think she thought of me as an election person, which I’m not,” Frank told the Inquirer somewhat incongruously.

Frank apparently did other things for Barnette, too. In an interview last year, he claimed to have been hired by a losing congressional campaign to knock on doors where his purported fake voters lived, in hopes of proving no legitimate voter was there. This was Barnette, though the Inquirer credits this door-knocking plan to another consultant.

The upshot?

“Barnette and a team of about 100 volunteers spend four days in December knocking on doors,” the paper’s Andrew Seidman reported. “About a third of the homes they visited ‘had something awry going on,’ Barnette said.” That’s what Frank said in that interview, too, that “we found about 30-32 percent of the doors we knocked on had at least one phantom voter.”

Now, it’s important to recognize why this is dubious. The set of houses being contacted was about 1,600, presumably all in the same county. Even with 100 people working for four days, you are not going to be able to confirm the identity of 1,600 residents, as anyone who’s ever done a door-knocking effort can attest. People aren’t home. People move. People don’t answer the door to strangers. It’s harder than it sounds. So we don’t know what the denominator of that 30 percent figure is. We do know, though, that there does not appear to have been any law enforcement action that identified numerous fraudulent votes in Montgomery County.

Despite that, Frank soon became a minor celebrity on the election-fraud circuit. He spoke at a Trump rally in Ohio, where he applied his magical polynomial to the state’s results to find the same “evidence of fraud,” by which I mean he discovered that turnout patterns were consistent there, too. He sat at MyPillow CEO Mike Lindell’s side as Lindell hosted a gathering of enthusiasts in South Dakota where he pledged to prove rampant fraud had occurred, ultimately failing to do so.

How did Lindell meet Frank? According to Seidman, through Barnette. She “helped introduce” the two, he wrote, with Frank’s claims quickly finding a home in Lindell’s pantheon of unverified assertions. “Lindell also interviewed Frank for a video in which they discuss Pennsylvania’s 4th District,” Seidman added.

Questions sent to Barnette’s campaign about her assessment of her 2020 loss and her work with Frank did not receive a response. (To Seidman, she said she “didn’t believe” the state’s elections were riddled with fraud.) The pair’s advocacy extended to presentations before state legislators and discussions of a lawsuit, an effort that ended when Barnette decided to run for Senate.

Frank and Lindell, though, are still at it. The pair spoke at an event in Gettysburg, Pa., last month, where they focused on “fixing the election.” Frank is the keynote speaker at an event including several right-wing candidates in Arizona this upcoming weekend. On a promotional image for the event, he is described as a “40 Year Modeler of Elections,” despite his insistence that he was not an “election person.”

On Sunday, Frank shared an update on his Telegram page: He and Lindell were endorsing Barnette’s Senate bid. If she comes up short in the hotly contested race, one can assume Frank will be able to explain why.