The Washington PostDemocracy Dies in Darkness

Ballot collectors are ‘mules.’ Skeptical reporters are ‘terrorists.’

True the Vote’s Gregg Phillips now says most of those appearing in ‘2000 Mules’ weren’t mules at all

Voters deliver their ballots to a polling station in Tempe, Ariz., on Nov. 3, 2020. (Matt York/AP)
8 min

One can certainly understand why Gregg Phillips of the conservative organization True the Vote might be feeling frustrated by the media.

Since his organization stepped into the spotlight as the engine behind Dinesh D’Souza’s film “2000 Mules” — a movie that uses geolocation data analyzed by Phillips to claim rampant ballot-gathering efforts in multiple swing states — there have been a number of stories detailing the holes or misrepresentations in the case Phillips hopes to make: that the data can’t actually prove what he and D’Souza suggest it does, for example, or that the purported data they show in the movie is obviously (and admittedly) fake.

His being in the spotlight also means that the media is obligated to point out how Phillips in 2016 made similar sweeping claims about illegal voting but completely failed to back up that assertion. More than enough to make a guy cranky.

But this is not just any guy. It’s Gregg Phillips, the guy who, in D’Souza’s film and at hearings in front of friendly state legislatures, has used intentionally pejorative language to cast his perceived opponents as criminals or dangerous. The people carrying ballots — which, if it occurred, wasn’t even illegal in two of the states True the Vote and D’Souza looked at — are “mules,” as in drug mules. The places where they allegedly collected these ballots (a part of the process left completely unproven in the film) become “stash houses.” All of it, he has declared, showed “organized crime.” It’s not subtle.

During a hearing before legislators in Arizona on Tuesday, Phillips applied that same tactic to members of the media.

“I think somebody wrote — the New York Times or someone — wrote, well, you can tell that they’re walking down the street, but you can’t see which side of the street they’re on,” he said. “Utter nonsense. It’s complete nonsense. They’re asking the wrong questions. They’re asking the wrong people.”

“These articles are being written by journalistic terrorists,” he added.

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

“Repeat that last line?” a legislator asked — drawing it out for emphasis, as she later made clear.

“Yeah,” Phillips said. “The articles are being written by journalistic terrorists, of which I think are probably one or two in this room.”

Kelli Ward, chair of the Republican Party in the state, liked that phrase.

In addition to retweeting Ward, the state party went further, declaring that the “mainstream media is Domestic Terrorists.” It later deleted that and another similar tweet.

What’s remarkable about Phillips’s comment isn’t simply how astonishingly over-the-top it was, or even how it made obvious his predilection for exaggerating nefariousness. What’s remarkable is that the new presentation did little to nothing to actually address the valid questions about his purported research. In fact, he made a comment that significantly weakened his case.

At issue is the reliability of his analysis of data collected from cellphones. The idea that thousands of people were ferrying ballots around battleground states depends on the regularity of those phones being in proximity to places with ballot drop boxes. Phillips and his supporters have regularly tried to suggest that people who doubt his analysis are casting doubt on the ability of phones to precisely geolocate someone, which is not the case. But it is the case that there’s good reason to think that Phillips’s analyses don’t include precise measurements of proximity to ballot drop boxes.

Consider what actually determining that someone visited a drop box would require. You’d need a precise measurement of location from the phone and you’d need to know precisely where the drop box was. In other words, you’d need to know that the phone was within a foot or two of a space that itself is only a few square feet. Not only will experts tell you that most cellphone geolocation data is not sufficiently precise, there’s no indication that the True the Vote people even knew where drop boxes were located at the libraries and government offices that hosted them.

The best indication we have of what True the Vote was looking at comes from the Georgia Bureau of Investigation, which in September determined that the data it was provided was insufficient to launch an investigation. The GBI said that the data was “cell site location information” — a triangulated location estimate less accurate than GPS — and that it centered on identifying visits “within 100 feet of a voter drop box.”

That’s significantly less precise than Phillips likes to suggest. During his testimony in Arizona, in fact, Phillips insisted that the military could get GPS precision down to a few centimeters. And that’s true — using specific technology that’s not generally available commercially. Using technology that there’s no reason to think he had. Instead, he insisted that by overlapping different bits of data — WiFi access and Bluetooth — he could get more precise. That’s not what the GBI says it received.

At one point, True the Vote’s team showed this map, purporting to show one person’s movements in Yuma County, Ariz.

“This is a sample pattern of life for one device in Yuma County from October 7th through November 3rd,” Phillips’s colleague Catherine Engelbrecht said. “This device went to 77 unique drop boxes. And you can see in the lower right where those locations were.”

Beyond all of the other questions that naturally arise — like, why bother going to different drop boxes, particularly in less-populated areas? — the data appear to show cellphone pings of someone who lives in Yuma (at top center) and occasionally drives around the region. Do those blue dots correspond to clear visits to libraries? Or more accurately, to visits to drop boxes at those libraries? Not to mention that Yuma County’s 2020 voter guide doesn’t include locations like the Foothills Library (at 13226 S. Frontage Rd. in Yuma) on its list of drop-off locations. We’re just meant to take Phillips’s and Engelbrecht’s word for it.

That’s what D’Souza did. I spoke to him about the movie and about Phillips’s data, repeatedly pressing him on questions about credibility. I asked him, for example, how he could be confident that people shown in the movie dropping off ballots were actually “mules,” given that there was no evidence provided for their having gone to more than one drop box. Considering one example of a guy on a bike dropping off a ballot, I pointed out to D’Souza that, beyond Phillips suggesting that the guy was being paid to deposit the ballot, the movie makes no direct claim and offers no evidence that he’s a “mule.”

“Of course there is a claim being made!” D’Souza insisted. “The context of the movie is to explain that that’s the search criterion. All the mules being counted under that rubric went to 10 or more drop boxes.”

Fine. But in Arizona, Phillips made a different claim — to again disparage journalists.

“One of the fallacies in people evaluating and fact-checking the movie,” he said, “really centered around a couple of videos that were out there of people shoving a whole bunch of ballots in.” That would presumably include the guy on the bike, who Phillips claimed (though it’s not apparent in the video) pulls “ballots” out of his backpack.

“Interestingly, only one of those people actually was a mule,” he continued, “because they didn’t meet the standard of having to get inside 10 times and go to one of the [nonprofit organizations] five times.”

You got that? Fact checks pointing out that the people included in the film “2000 Mules” were not actually mules were flawed because only one of the people shown in the film actually met Phillips’s “mule” standard anyway. (An email to Phillips asking him to identify which person was the mule did not receive a response by the time of publication though he later wrote to say the quote was “out of context.”)

The person shown in the movie before the guy on the bike, incidentally, was a woman who Engelbrecht claimed had visited “dozens and dozens” of drop boxes in both the 2020 presidential election and the Georgia Senate runoff in January 2021. So is she the sole mule? Or was Engelbrecht offering another misrepresentation to viewers? And if so, did D’Souza know?

It’s because of things like this that reporters urge caution in viewing the movie or in taking Phillips’s claims at face value. And it’s because we urge caution that Phillips compares us to people who use violence to evoke fear.