The Washington PostDemocracy Dies in Darkness

Let’s be very clear about Douglas Frank’s allegations of voter fraud

Douglas Frank, before former president Donald Trump's arrival at a June rally in Wellington, Ohio. (Jabin Botsford/The Washington Post)

In a lengthy story published Wednesday, The Washington Post’s Amy Gardner, Emma Brown and Josh Dawsey explored the ongoing effort by allies of former president Donald Trump to prove the unprovable assertion that the 2020 presidential election was stolen. It is unprovable for the simple reason that the election wasn’t stolen. It is as unprovable as the idea that the sun is actually made out of popcorn.

If there was attention and income to be generated from claiming that the sun was made of popcorn, though, you can be sure that there would soon emerge a galaxy of people intent on making just such an assertion. Follow them on Twitter and pay for their symposia and you, too, can learn the secrets behind the sun’s buttery goodness.

In The Post’s report, one of the foremost purveyors of the claims of election fraud is a high school teacher named Douglas Frank. Gardner, Brown and Dawsey detail Frank’s conversations with state officials and his online advocacy — statements that at times spin into dangerous territory.

But it’s also worth pointing out that Frank’s claims have been shown to be nonsensical. Repeatedly. By us.

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

Consider this quote, from North Carolina state Rep. George Cleveland (R).

“When I first saw Frank’s presentation on the polynomial thing, I was impressed,” Cleveland said, referring to the claim that an algorithm Frank calls a “6th-degree polynomial” was used to rig the 2020 results. “He did a very good job. But after I’ve looked into some other things and talked to other people, until he will give his data and his methodology to others to prove what he’s saying, I’m going to be very skeptical.”

Mr. Cleveland, your skepticism will surely be rewarded.

We’ve been over this before, but it bears reiterating. Frank’s sixth-degree polynomial is a bit of impressive-sounding chicanery that is light-years away from any proof of fraud.

The way it works is that Frank takes turnout data by age from several counties in a state. He then averages the turnout in those counties — the percent of 18-year-olds that voted, the percent of 33-year-olds, etc. Then he drops that into Excel and uses its Trendline tool to create that “sixth-degree polynomial,” basically a curve that’s designed to fit as closely to the data points as Excel will allow. Then he compares all the other counties to that line and — completely unsurprisingly — discovers that the turnout in all of the counties is pretty close to that curve.

I can’t believe that this still needs to be articulated months after Frank first made this claim, but this is idiotic. It is idiotic because there are consistent age-related turnout patterns across the country. Older people vote more; younger people vote less. In states, counties tend to follow that pattern. Here’s what turnout looked like in Ohio, for example, using a graph of data from the political data firm L2 from when I debunked Frank’s claims back in June. That solid line is the average of turnout by age across all counties. And guess what! Each county’s turnout is awfully close to that average. Because that’s how turnout consistently works and that’s how averages generally work.

I like to use a simpler analogy. Imagine claiming that Usain Bolt was cheating because you’d averaged the results of five of his races and discovered that all of his results were close to that average. That’s literally the claim that Frank is making — that the closeness to the average proves that fraud occurred. It’s ridiculous.

Frank has other components to his hustle, too. At the top of The Post’s story dealing with him is a slide from a slide show he presented at a rally in Ohio. You can view a similar version online, if you want. It doesn’t prove any actual fraud, just makes claims such as there were more people registered to vote in some states than were eligible to vote in those states. This is because Frank doesn’t understand how census data works or that estimates of population do not actually precisely measure the number of people in a state. There’s never actual evidence of fraud or illegal registrations, just purported evidence of evidence of shenanigans.

At other times, Frank has claimed to have spoken with voters who were recorded as having cast votes but who now say they didn’t. He’s never brought any such voter forward, instead making broad claims about door-to-door canvassing that anyone who has spent more than an hour trying to reach a determined list of individuals at home will immediately recognize as questionable. Frank has, however, made claims about absentee ballots in Ohio and voters in Alabama that officials in those states quickly debunked (as documented in The Post’s report).

I reiterate this because I have a strong affection for both math and the truth. That someone could make data-based claims for months after they’ve been obviously shown to be nonsense is something that I consider a personal affront. I understand that a lot of people want to believe that the sun is made out of popcorn. But that doesn’t mean we should simply let bad actors try to use complicated-sounding nonsense to convince them that it is.

The hero of our story about Frank’s claims, at least in my eyes, is Ohio Secretary of State Frank LaRose (R). Speaking to The Post’s reporters, LaRose made the very good point that terms like “sixth-order polynomial” are inscrutable to most people. It’s like “Star Trek” gibberish — engage the quantum polyspheric hyperdrive! — offered in service of undermining the results of the election.

“It sounds impressive,” LaRose said of the “polynomial” thing, “and if you have a preconceived sort of notion about what you want to believe and somebody with a title tells you a bunch of things, this is what conspiracy theories are based on.”

Precisely right.