The Trump Voter-Fraud Commission's Data Problem

During its second meeting on Tuesday, the Kris Kobach-led group heard presentations that sometimes followed a circular logic.

Holly Ramer / AP

“Election Fraud is rampant!! California has 11 Counties that have MORE VOTERS than registered voters!!” reads one of the most recent public comments submitted to the Presidential Advisory Commission on Election Integrity. “I have personally witnessed voter fraud in California,” reads another comment. An earlier comment claims that “many [voters] were deceased and many were not citizens.”  Yet another post is from a former San Diego poll watcher who claims to have witnessed attempted voter fraud and was told by an elections official: “if someone wants to vote I am not about to stop them. This is America, not China!”

Over 500 comments were submitted online during the run-up to the second meeting of the commission on Tuesday, which was led by its vice-chair Kansas Secretary of State Kris Kobach. Many of those comments were in the same vein as those from the San Diego poll watcher, containing anecdotes or hearsay about egregious incidents of voter fraud or alleging massive levels of fraud on a national level. While those public comments weren’t directly addressed during the proceedings of the meeting, they did help set its tone, as Kobach and his fellow commissioners grappled with the gap between rhetoric on voter fraud—backed by ample anecdote—and data on in-person voter fraud, which are scant.

The meeting started with a presentation from Andrew Smith, a political scientist at the University of New Hampshire, that addressed data challenges in studying voter behavior, elections laws, and turnout. His presentation raised questions about how to interpret elections data and registration data (especially when multiple databases with different population denominators are in use), and the problems with correlating voter-ID policies with decreases in turnout. Smith’s warnings about the fuzziness of elections data made sense, but his data on turnout conflicted with some of the findings of the Government Accountability Office, which found that three of five studies to meet its standard of rigor showed that voter-ID laws specifically decreased turnout of people of color relative to whites.

The opening discussion of data, turnout, and policy mattered deeply to the commission, which was chartered by President Trump in May, and has been dogged by allegations that its true purpose is not eliminating voter fraud, but instigating voter suppression. The argument in favor of the legitimacy of the commission has been weakened by the tendency of Trump and Kobach to commit the logical fallacy of begging the question: They have invoked unsubstantiated claims of widespread voter fraud in order to argue for the necessity of a commission to prove that voter fraud exists. And that commission so far—while ostensibly still waiting to analyze numbers from a controversial set of voter data gathered from the states—has simultaneously argued in favor of measures to fight voter fraud, like voter ID, while also arguing that policies to do so won’t suppress votes.

Several of the commissioners and invited guests used fuzzy or opaque data claims not just to further requests to study voter fraud, but to also push anti-voter-fraud legislation. Ken Block, the president of a software firm that used a consumer dataset to match the names and Social Security numbers of voters, claimed that his group was “able to identify with high confidence, several examples of voter fraud” and that “no government agency is looking for voter fraud.” * The accuracy of the latter statement aside—states certify elections results and check for double voters using much more complete datasets than his—Block never detailed the data collection methodology of his proprietary data, which a report shows come from Virtual DBS, a consumer data firm. ** If that dataset contains false duplicates or imputes the wrong Social Security number to two people with identical names, it would create duplicate voters for Block to pick up.

Block’s data was leaned on heavily by other commissioners, including the Heritage Foundation fellow and former elections official Hans von Spakovsky, a grizzled veteran of the quest for the grand voter conspiracy, who also pointed out several other unverified reports of double voting and noncitizen voting, including a group of about 300 people he’d flagged for registering to vote as noncitizens in Fairfax, Virginia, although no investigation materialized. Von Spakovsky expressed his conviction that there are “hundreds, if not thousands of voter-fraud cases that have yet to be investigated.”

During the meeting, even  Kobach faced the wide chasm between his own claims of widespread voter-fraud and the evidence. On September 7, Kobach—who has launched hundreds of investigations of voter fraud in his own state and walked away with fewer than a dozen convictions—claimed in a column he wrote for Breitbart (where he is a paid contributor) that “facts have come to light that indicate that a pivotal, close election was likely changed through voter fraud on November 8, 2016: New Hampshire’s U.S. Senate Seat, and perhaps also New Hampshire’s four electoral college votes in the presidential election.” To Kobach, the source of that fraud was mostly thousands of out-of-staters who didn’t have updated driver’s licenses, and thus “never were bona fide residents of the State.”

But, on Tuesday, when confronted with data directly challenging his claim—specifically, New Hampshire only requires voters to be domiciled, not residents, which means that out-of-staters can vote legitimately—Kobach changed his claim, instead criticizing the New Hampshire law, even as New Hampshire Secretary of State Bill Gardner—who is also a commissioner—called Kobach’s column “reckless.” Still, much of the meeting and its segment on voter fraud focused on New Hampshire’s voting laws, a focus brought on in part by Kobach’s own disupted column.

That episode reflects the nature of Trump’s voter-fraud commission and the modus operandi of many politicians who claim voter fraud is rampant: They make false, unverified, or disputed claims of fraudulent voting, then use those claims as evidence that anti-voter-fraud efforts are necessary, and when claims are debunked, cite that as a demonstration of the need to pursue the subject more thoroughly. No matter how comprehensive the data, both analyses that purport to document the presence of voter fraud and those that are taken as signals of the need to dig deeper and pass further legislation.

That logic undercuts what could be an invaluable effort at improving turnout, modernizing registrations and voter rolls, improving elections technology, increasing cross-state communication, reducing administrative problems, and pursuing discrepancies and instances of illegal elections actions that are prevalent and well known. Those actions are included in the charges of Kobach’s commission, though it might not have been evident in the way they were squeezed into the last panel discussion of the day, nor in the overwhelming focus of the public comments on fraud and suppression.

*This article originally stated that Block's firm matched partial Social Security numbers instead of full ones. We regret the error.

** This article has been updated to clarify the source of proprietary data.