In the end, the decision seemed inevitable. After a seven-day trial in Kansas City federal court in March, in which Kansas Secretary of State Kris Kobach needed to be tutored on basic trial procedure by the judge and was found in contempt for his “willful failure” to obey a ruling, even he knew his chances were slim. Kobach told The Kansas City Star at the time that he expected the judge would rule against him (though he expressed optimism in his chances on appeal).

Sure enough, yesterday federal Judge Julie Robinson overturned the law that Kobach was defending as lead counsel for the state, dealing him an unalloyed defeat. The statute, championed by Kobach and signed into law in 2013, required Kansans to present proof of citizenship in order to register to vote. The American Civil Liberties Union sued, contending that the law violated the National Voter Registration Act (AKA the “motor voter” law), which was designed to make it easy to register.

The trial had a significance that extends far beyond the Jayhawk state. One of the fundamental questions in the debate over alleged voter fraud — whether a substantial number of non-citizens are in fact registering to vote — was one of two issues to be determined in the Kansas proceedings. (The second was whether there was a less burdensome solution than what Kansas had adopted.) That made the trial a telling opportunity to remove the voter fraud claims from the charged, and largely proof-free, realms of political campaigns and cable news shoutfests and examine them under the exacting strictures of the rules of evidence.

That’s precisely what occurred and according to Robinson, an appointee of George W. Bush, the proof that voter fraud is widespread was utterly lacking. As the judge put it, “the court finds no credible evidence that a substantial number of noncitizens registered to vote” even under the previous law, which Kobach had claimed was weak.

For Kobach, the trial should’ve been a moment of glory. He’s been arguing for a decade that voter fraud is a national calamity. Much of his career has been built on this issue, along with his fervent opposition to illegal immigration. (His claim is that unlawful immigrants are precisely the ones voting illegally.) Kobach, who also co-chaired the Trump administration’s short-lived commission on voter fraud, is perhaps the individual most identified with the cause of sniffing out and eradicating phony voter registration. He’s got a gilded resume, with degrees from Harvard University, Yale Law School and the University of Oxford, and is seen as both the intellect behind the cause and its prime advocate. Kobach has written voter laws in other jurisdictions and defended them in court. If anybody ever had time to marshal facts and arguments before a trial, it was Kobach.

But things didn’t go well for him in the Kansas City courtroom, as Robinson’s opinion made clear. Kobach’s strongest evidence of non-citizen registration was anemic at best: Over a 20-year period, fewer than 40 non-citizens had attempted to register in one Kansas county that had 130,000 voters. Most of those 40 improper registrations were the result of mistakes or confusion rather than intentional attempts to mislead, and only five of the 40 managed to cast a vote.

One of Kobach’s own experts even rebutted arguments made by both Kobach and President Donald Trump. The expert testified that a handful of improper registrations could not be extrapolated to conclude that 2.8 million fraudulent votes — roughly, the gap between Hillary Clinton and Trump in the popular vote tally — had been cast in the 2016 presidential election. Testimony from a second key expert for Kobach also fizzled.

As the judge’s opinion noted, Kobach insisted the meager instances of cheating revealed at trial are just “the tip of the iceberg.” As she explained, “This trial was his opportunity to produce credible evidence of that iceberg, but he failed to do so.” Dismissing the testimony by Kobach’s witnesses as unpersuasive, Robinson drew what she called “the more obvious conclusion that there is no iceberg; only an icicle largely created by confusion and administrative error.”

By the time the trial was over, Kobach, a charismatic 52-year-old whose broad shoulders and imposing height make him resemble an aging quarterback, seemed to have shrunk inside his chair at the defense table.

But despite his defeat, Kobach’s causes — restricting immigration and tightening voting requirements — seem to be enjoying favorable tides elsewhere. Recent press accounts noted Kobach’s role in restoring a question about citizenship, abandoned since 1950, to U.S. Census forms for 2020. And the Supreme Court ruled on June 11 that the state of Ohio can purge voters from its rolls when they fail to vote even a single time and don’t return a mailing verifying their address, a provision that means more voters will need to re-register and prove their eligibility again.

For his own part, Kobach is now a candidate for governor of Kansas, running neck and neck with the incumbent in polls for the Republican primary on Aug. 7. It’s not clear whether the verdict will affect his chances — or whether it will lead him and others to quietly retreat from claims of voter fraud. But the judge’s opinion and expert interviews reveal that Kobach effectively put the concept of mass voter fraud to the test — and the evidence crumbled.

Perhaps it was an omen. Before Kobach could enter the courtroom inside the Robert J. Dole U.S. Courthouse each day, he had to pass through a hallway whose walls featured a celebratory display entitled “Americans by Choice: The Story of Immigration and Citizenship in Kansas.” Photographs of people who’d been sworn in as citizens in that very courthouse were superimposed on the translucent window shades.

Public interest in the trial was high. The seating area quickly filled to capacity on the first day of trial on the frigid morning of March 6. The jury box was opened to spectators; it wouldn’t be needed, as this was a bench trial. Those who couldn’t squeeze in were sent to a lower floor, where a live feed had been prepared in a spillover room.

From the moment the trial opened, Kobach and his co-counsels in the Kansas secretary of state’s office, Sue Becker and Garrett Roe, stumbled over the most basic trial procedures. Their mistakes antagonized the judge. “Evidence 101,” Robinson snapped, only minutes into the day, after Kobach’s team attempted to improperly introduce evidence. “I’m not going to do it.”

Matters didn’t improve for Kobach from there.

Throughout the trial, his team’s repeated mishaps and botched cross examinations cost hours of the court’s time. Robinson was repeatedly forced to step into the role of law professor, guiding Kobach, Becker and Roe through courtroom procedure. “Do you know how to do the next step, if that’s what you’re going to do?” the judge asked Becker at one point, as she helped her through the steps of impeaching a witness. “We’re going to follow the rules of evidence here.”  

Becker often seemed nervous. She took her bright red glasses off and on. At times she burst into nervous chuckles after a misstep. She laughed at witnesses, skirmished with the judge and even taunted the lawyers for the ACLU. “I can’t wait to ask my questions on Monday!” she shouted at the end of the first week, jabbing a finger in the direction of Dale Ho, the lead attorney for the plaintiffs. Ho rolled his eyes.

Roe was gentler — deferential, even. He often admitted he didn’t know what step came next, asking the judge for help. “I don’t — I don’t know if this one is objectionable. I hope it’s not,” he offered at one point, as he prepared to ask a question following a torrent of sustained objections. “I’ll let you know,” an attorney for the plaintiffs responded, to a wave of giggles in the courtroom. On the final day of trial, as Becker engaged in yet another dispute with the judge, Roe slapped a binder to his forehead and audibly whispered, “Stop talking. Stop talking.”

Kobach’s cross examinations were smoother and better organized, but he regularly attempted to introduce exhibits — for example, updated state statistics that he had failed to provide the ACLU in advance to vet — that Robinson ruled were inadmissible. As the trial wore on, she became increasingly irritated. She implored Kobach to “please read” the rules on which she based her rulings, saying his team had repeated these errors “ad nauseum.”

Kobach seemed unruffled. Instead of heeding her advice, he’d proffer the evidence for the record, a practice that allows the evidence to be preserved for appeal even if the trial judge refuses to admit it. Over the course of the trial, Kobach and his team would do this nearly a dozen times.

Eventually, Robinson got fed up. She asked Kobach to justify his use of proffers. Kobach, seemingly alarmed, grabbed a copy of the Federal Rules of Civil Procedure — to which he had attached a growing number of Post-it notes — and quickly flipped through it, trying to find the relevant rule.

The judge tried to help. “It’s Rule 26, of course, that’s been the basis for my rulings,” she told Kobach. “I think it would be helpful if you would just articulate under what provision of Rule 26 you think this is permissible.” Kobach seemed to play for time, asking clarifying questions rather than articulating a rationale. Finally, the judge offered mercy: a 15-minute break. Kobach’s team rushed from the courtroom.

It wasn’t enough to save him. In her opinion, Robinson described “a pattern and practice by Defendant [Kobach] of flaunting disclosure and discovery rules.” As she put it, “it is not clear to the Court whether Defendant repeatedly failed to meet his disclosure obligations intentionally or due to his unfamiliarity with the federal rules.” She ordered Kobach to attend the equivalent of after-school tutoring: six hours of extra legal education on the rules of civil procedure or the rules of evidence (and to present the court with a certificate of completion).

It’s always a bad idea for a lawyer to try the patience of a judge — and that’s doubly true during a bench trial, when the judge will decide not only the law, but also the facts. Kobach repeatedly annoyed Robinson with his procedural mistakes. But that was nothing next to what the judge viewed as Kobach’s intentional bad faith.

This view emerged in writing right after the trial — that’s when Robinson issued her ruling finding Kobach in contempt — but before the verdict. And the conduct that inspired the contempt finding had persisted over several years. Robinson concluded that Kobach had intentionally failed to follow a ruling she issued in 2016 that ordered him to restore the privileges of 17,000 suspended Kansas voters.

In her contempt ruling, the judge cited Kobach’s “history of noncompliance” with the order and characterized his explanations for not abiding by it as “nonsensical” and “disingenuous.” She wrote that she was “troubled” by Kobach’s “failure to take responsibility for violating this Court’s orders, and for failing to ensure compliance over an issue that he explicitly represented to the Court had been accomplished.” Robinson ordered Kobach to pay the ACLU’s legal fees for the contempt proceeding.

That contempt ruling was actually the second time Kobach was singled out for punishment in the case. Before the trial, a federal magistrate judge deputized to oversee the discovery portion of the suit fined him $1,000 for making “patently misleading representations” about a voting fraud document Kobach had prepared for Trump. Kobach paid the fine with a state credit card.

More than any procedural bumbling, the collapse of Kobach’s case traced back to the disintegration of a single witness.

The witness was Jesse Richman, a political scientist from Old Dominion University, who has written studies on voter fraud. For this trial, Richman was paid $5,000 by the taxpayers of Kansas to measure non-citizen registration in the state. Richman was the man who had to deliver the goods for Kobach.

With his gray-flecked beard and mustache, Richman looked the part of an academic, albeit one who seemed a bit too tall for his suit and who showed his discomfort in a series of awkward, sudden movements on the witness stand. At moments, Richman’s testimony turned combative, devolving into something resembling an episode of The Jerry Springer Show. By the time he left the stand, Richman had testified for more than five punishing hours. He’d bickered with the ACLU’s lawyer, raised his voice as he defended his studies and repeatedly sparred with the judge.

“Wait, wait, wait!” shouted Robinson at one point, silencing a verbal free-for-all that had erupted among Richman, the ACLU’s Ho, and Kobach, who were all speaking at the same time. “Especially you,” she said, turning her stare to Richman. “You are not here to be an advocate. You are not here to trash the plaintiff. And you are not here to argue with me.”

Richman had played a small but significant part in the 2016 presidential campaign. Trump and others had cited his work to claim that illegal votes had robbed Trump of the popular vote. At an October 2016 rally in Wisconsin, the candidate cited Richman’s work to bolster his predictions that the election would be rigged. “You don’t read about this, right?” Trump told the crowd, before reading from an op-ed Richman had written for The Washington Post: “‘We find that this participation was large enough to plausibly account for Democratic victories in various close elections.’ Okay? All right?”

Richman’s 2014 study of non-citizen registration used data from the Cooperative Congressional Election Study — an online survey of more than 32,000 people. Of those, fewer than 40 individuals indicated they were non-citizens registered to vote. Based on that sample, Richman concluded that up to 2.8 million illegal votes had been cast in 2008 by non-citizens. In fact, he put the illegal votes at somewhere between 38,000 and 2.8 million — a preposterously large range — and then Trump and others simply used the highest figure.

Academics pilloried Richman’s conclusions. Two hundred political scientists signed an open letter criticizing the study, saying it should “not be cited or used in any debate over fraudulent voting.” Harvard’s Stephen Ansolabehere, who administered the CCES, published his own peer-reviewed paper lambasting Richman’s work. Indeed, by the time Trump read Richman’s article onstage in 2016, The Washington Post had already appended a note to the op-ed linking to three rebuttals and a peer-reviewed study debunking the research.

None of that discouraged Kobach or Trump from repeating Richman’s conclusions. They then went a few steps further. They took the top end of the range for the 2008 election, assumed that it applied to the 2016 election, too, and further assumed that all of the fraudulent ballots had been cast for Clinton.

Some of those statements found their way into the courtroom, when Ho pressed play on a video shot by The Kansas City Star on Nov. 30, 2016. Kobach had met with Trump 10 days earlier and had brought with him a paper decrying non-citizen registration and voter fraud. Two days later, Trump tweeted that he would have won the popular vote if not for “millions of people who voted illegally.”

On the courtroom’s televisions, Kobach appeared, saying Trump’s tweet was “absolutely correct.” Without naming Richman, Kobach referred to his study: The number of non-citizens who said they’d voted in 2008 was far larger than the popular vote margin, Kobach said on the video. The same number likely voted again in 2016.

In the courtroom, Ho asked Richman if he believed his research supported such a claim. Richman stammered. He repeatedly looked at Kobach, seemingly searching for a way out. Ho persisted and finally, Richman gave his answer: “I do not believe my study provides strong support for that notion.”

To estimate the number of non-citizens voting in Kansas, Richman had used the same methodology he employed in his much-criticized 2014 study. Using samples as small as a single voter, he’d produced surveys with wildly different estimates of non-citizen registration in the state. The multiple iterations confused everyone in the courtroom.

“For the record, how many different data sources have you provided?” Robinson interjected in the middle of one Richman answer. “You provide a range of, like, zero to 18,000 or more.”

“I sense the frustration,” Richman responded, before offering a winding explanation of the multiple data sources and surveys he’d used to arrive at a half-dozen different estimates. Robinson cut him off. “Maybe we need to stop here,” she said.

“Your honor, let me finish answering your question,” he said.

“No, no. I’m done,” she responded, as he continued to protest. “No. Dr. Richman, I’m done.”

To refute Richman’s numbers, the ACLU called on Harvard’s Ansolabehere, whose data Richman had relied on in the past. Ansolabehere testified that Richman’s sample sizes were so small that it was just as possible that there were no non-citizens registered to vote in Kansas as 18,000. “There’s just a great deal of uncertainty with these estimates,” he said.

Ho asked if it would be accurate to say that Richman’s data “shows a rate of non-citizen registration in Kansas that is not statistically distinct from zero?”

“Correct.”

The judge was harsher than Ansolabehere in her description of Richman’s testimony. In her opinion, Robinson unloaded a fusillade of dismissive adjectives, calling Richman’s conclusions “confusing, inconsistent and methodologically flawed,” and adding that they were “credibly dismantled” by Ansolabehere. She labeled elements of Richman’s testimony “disingenuous” and “misleading,” and stated that she gave his research “no weight” in her decision.

One of the paradoxes of Kobach is that he has become a star in circles that focus on illegal immigration and voting fraud despite poor results in the courtroom. By ProPublica’s count, Kobach chalked up a 2–6 won-lost record in federal cases in which he was played a major role, and which reached a final disposition before the Kansas case.

Those results occurred when Kobach was an attorney for the legal arm of the Federation for American Immigration Reform from 2004 to 2011, when he became secretary of state in Kansas. In his FAIR role (in which he continued to moonlight till about 2014), Kobach traveled to places like Fremont, Nebraska, Hazleton, Pennsylvania, Farmers Branch, Texas, and Valley Park, Missouri, to help local governments write laws that attempted to hamper illegal immigration, and then defend them in court. Kobach won in Nebraska, but lost in Texas and Pennsylvania, and only a watered down version of the law remains in Missouri.

The best-known law that Kobach helped shape before joining the Kansas government in 2011 was Arizona’s “show me your papers” law. That statute allowed police to demand citizenship documents for any reason from anyone they thought might be in the country illegally. After it passed, the state paid Kobach $300 an hour to train law enforcement on how to legally arrest suspected illegal immigrants. The Supreme Court gutted key provisions of the law in 2012.

Kobach also struggled in two forays into political campaigning. In 2004, he lost a race for Congress. He also drew criticism for his stint as an informal adviser to Mitt Romney’s 2012 presidential campaign. Kobach was the man responsible for Romney’s much-maligned proposal that illegal immigrants “self-deport,” one reason Romney attracted little support among Latinos. Romney disavowed Kobach even before the campaign was over, telling media outlets that he was a “supporter,” not an adviser.

Trump’s election meant Kobach’s positions on immigration would be welcome in the White House. Kobach lobbied for, but didn’t receive, an appointment as Secretary of Homeland Security. He was, however, placed in charge of the voter fraud commission, a pet project of Trump’s. Facing a raft of lawsuits and bad publicity, the commission was disbanded little more than six months after it formally launched.

Back at home, Kobach expanded his power as secretary of state. Boasting of his experience as a law professor and scholar, Kobach convinced the state legislature to give him the authority to prosecute election crimes himself, a power wielded by no other secretary of state. In that role, he has obtained nine guilty pleas against individuals for election-related misdemeanors. Only one of those who pleaded guilty, as it happens, was a non-citizen.

He also persuaded Kansas’ attorney general to allow Kobach to represent the state in the trial of Kansas’ voting law. Kobach argued it was a bargain. As he told The Wichita Eagle at the time, “The advantage is the state gets an experienced appellate litigator who is a specialist in this field and in constitutional law for the cost the state is already paying, which is my salary.”

Kobach fared no better in the second main area of the Kansas City trial than he had in the first. This part explored whether there is a less burdensome way of identifying non-citizens than forcing everyone to show proof of citizenship upon registration. Judge Robinson would conclude that there were many alternatives that were less intrusive.

In his opening, Ho of the ACLU spotlighted a potentially less intrusive approach. Why not use the Department of Homeland Security’s Systematic Alien Verification for Entitlements System list, and compare the names on it to the Kansas voter rolls? That, Ho argued, could efficiently suss out illegal registrations.

Kobach told the judge that simply wasn’t feasible. The list, he explained, doesn’t contain all non-citizens in the country illegally — it contains only non-citizens legally present and those here illegally who register in some way with the federal government. Plus, he told Robinson, in order to really match the SAVE list against a voter roll, both datasets would have to contain alien registration numbers, the identifier given to non-citizens living in the U.S. “Those are things that a voter registration system doesn’t have,” he said. “So, the SAVE system does not work.”

But Kobach had made the opposite argument when he headed the voter fraud commission. There, he’d repeatedly advocated the use of the SAVE database. Appearing on Fox News in May 2017, shortly after the commission was established, Kobach said, “The Department of Homeland Security knows of the millions of aliens who are in the United States legally and that data that’s never been bounced against the state’s voter rolls to see whether these people are registered.” He said the federal databases “can be very valuable.”

A month later, as chief of the voting fraud commission, Kobach took steps to compare state information to the SAVE database. He sent a letter to all 50 secretaries of state requesting their voter rolls. Bipartisan outrage ensued. Democrats feared he would use the rolls to encourage states to purge legitimately registered voters. Republicans labelled the request federal overreach.

At trial, Kobach’s main expert on this point was Hans von Spakovsky, another member of the voter fraud commission. He, too, had been eager in commission meetings to match state voter rolls to the SAVE database.

But like Kobach, von Spakovsky took a different tack at trial. He testified that this database was unusable by elections offices. “In your experience and expertise as an election administrator and one who studies elections,” Kobach asked, “is [the alien registration number] a practical or even possible thing for a state to do in its voter registration database?” Von Spakovsky answered, “No, it is not.”

Von Spakovsky and Kobach have been friends for more than a decade. They worked together at the Department of Justice under George W. Bush. Kobach focused on immigration issues — helping create a database to register visitors to the U.S. from countries associated with terrorism — while von Spakovsky specialized in voting issues; he had opposed the renewal of the Voting Rights Act.

Von Spakovsky’s history as a local elections administrator in Fairfax County, Va., qualified him as an expert on voting fraud. Between 2010 and 2012, while serving as vice chairman of the county’s three-member electoral board, he’d examined the voter rolls and found what he said were 300 registered non-citizens. He’d pressed for action against them, but none came. Von Spakovsky later joined the Heritage Foundation, where he remains today, generating research that underpins the arguments of those who claim mass voter fraud.

Like Richman, von Spakovsky seemed nervous on the stand, albeit not combative. He wore wire-rimmed glasses and a severe, immovable expression. Immigration is a not-so-distant feature of his family history: His parents — Russian and German immigrants — met in a refugee camp in American-occupied Germany after World War II before moving to the U.S.

Von Spakovsky had the task of testifying about what was intended to be a key piece of evidence for Kobach’s case: a spreadsheet of 38 non-citizens who had registered to vote, or attempted to register, in a 20-year period in Sedgwick County, Kansas.

But the 38 non-citizens turned out to be something less than an electoral crime wave. For starters, some of the 38 had informed Sedgwick County that they were non-citizens. One woman had sent her registration postcard back to the county with an explanation that it was a “mistake” and that she was not a citizen. Another listed an alien registration number — which tellingly begins with an “A” — instead of a Social Security number on the voter registration form. The county registered her anyway.

When von Spakovsky took the stand, he had to contend with questions that suggested he had cherry-picked his data. (The judge would find he had.) In his expert report, von Spakovsky had referenced a 2005 report by the Government Accountability Office that polled federal courts to see how many non-citizens had been excused from jury duty for being non-citizens — a sign of fraud, because jurors are selected from voter rolls. The GAO report mentioned eight courts. Only one said it had a meaningful number of jury candidates who claimed to be non-citizens: “between 1 and 3 percent” had been dismissed on these grounds. This was the only court von Spakovsky mentioned in his expert report.

His report also cited a 2012 TV news segment from an NBC station in Fort Myers, Fla. Reporters claimed to have discovered more than 100 non-citizens on the local voter roll.

“Now, you know, Mr. von Spakovsky, don’t you, that after this NBC report there was a follow-up by the same NBC station that determined that at least 35 of those 100 individuals had documentation to prove they were, in fact, United States citizens. Correct?” Ho asked. “I am aware of that now, yes,” von Spakovsky replied.

That correction had been online since 2012 and Ho had asked von Spakovsky the same question almost two years before in a deposition before the trial. But von Spakovsky never corrected his expert report.

Under Ho’s questioning, von Spakovsky also acknowledged a false assertion he made in 2011. In a nationally syndicated column for McClatchy, von Spakovsky claimed a tight race in Missouri had been decided by the illegal votes of 50 Somali nationals. A month before the column was published, a Missouri state judge ruled that no such thing had happened.

On the stand, von Spakovsky claimed he had no knowledge of the ruling when he published the piece. He conceded that he never retracted the assertion.

Kobach, who watched the exchange without objection, had repeatedly made the same claim — even after the judge ruled it was false. In 2011, Kobach wrote a series of columns using the example as proof of the need for voter ID, publishing them in outlets ranging from the Topeka Capital-Journal to the Wall Street Journal and the Washington Post. In 2012, he made the claim in an article published in the Syracuse Law Review. In 2013, he wrote an op-ed for the Kansas City Star with the same example: “The election was stolen when Rizzo received about 50 votes illegally cast by citizens of Somalia.” None of those articles have ever been corrected.

Ultimately, Robinson would lacerate von Spakovsky’s testimony, much as she had Richman’s. Von Spakovsky’s statements, the judge wrote, were “premised on several misleading and unsupported examples” and included “false assertions.” As she put it, “His generalized opinions about the rates of noncitizen registration were likewise based on misleading evidence, and largely based on his preconceived beliefs about this issue, which has led to his aggressive public advocacy of stricter proof of citizenship laws.”

There was one other wobbly leg holding up the argument that voter fraud is rampant: the very meaning of the word “fraud.”

Kobach’s case, and the broader claim, rely on an extremely generous definition. Legal definitions of fraud require a person to knowingly be deceptive. But both Kobach and von Spakovsky characterized illegal ballots as “fraud” regardless of the intention of the voter.

Indeed, the nine convictions Kobach has obtained in Kansas are almost entirely made up of individuals who didn’t realize they were doing something wrong. For example, there were older voters who didn’t understand the restrictions and voted in multiple places they owned property. There was also a college student who’d forgotten she’d filled out an absentee ballot in her home state before voting months later in Kansas. (She voted for Trump both times.)

Late in the trial, the ACLU presented Lorraine Minnite, a professor at Rutgers who has written extensively about voter fraud, as a rebuttal witness. Her book, “The Myth of Voter Fraud,” concluded that almost all instances of illegal votes can be chalked up to misunderstandings and administrative error.

Kobach sent his co-counsel, Garrett Roe, to cross-examine her. “It’s your view that what matters is the voter’s knowledge that his or her action is unlawful?” Roe asked. “In a definition of fraud, yes,” said Minnite. Roe pressed her about this for several questions, seemingly surprised that she wouldn’t refer to all illegal voting as fraud.

Minnite stopped him. “The word ‘fraud’ has meaning, and that meaning is that there’s intent behind it. And that’s actually what Kansas laws are with respect to illegal voting,” she said. “You keep saying my definition” she said, putting finger quotes around “my.” “But, you know, it’s not like it’s a freak definition.”

Kobach had explored a similar line of inquiry with von Spakovsky, asking him if the list of 38 non-citizens he’d reviewed could be absolved of “fraud” because they may have lacked intent.

“No,” von Spakovsky replied, “I think any time a non-citizen registers, any time a non-citizen votes, they are — whether intentionally or by accident, I mean — they are defrauding legitimate citizens from a fair election.”

After Kobach concluded his questions, the judge began her own examination of von Spakovsky.

“I think it’s fair to say there’s a pretty good distinction in terms of how the two of you define fraud,” the judge said, explaining that Minnite focused on intent, while she understood von Spakovsky’s definition to include any time someone who wasn’t supposed to vote did so, regardless of reason. “Would that be a fair characterization?” she asked.

“Yes ma’am,” von Spakovsky replied.

The judge asked whether a greater number of legitimate voters would be barred from casting ballots under the law than fraudulent votes prevented. In that scenario, she asked, “Would that not also be defrauding the electoral process?” Von Spakovsky danced around the answer, asserting that one would need to answer that question in the context of the registration requirements, which he deemed reasonable.

The judge cut him off. “Well that doesn’t really answer my question,” she said, saying that she found it contradictory that he wanted to consider context when examining the burden of registration requirements, but not when examining the circumstances in which fraud was committed.

“When you’re talking about … non-citizen voting, you don’t want to consider that in context of whether that person made a mistake, whether a DMV person convinced them they should vote,” she said. Von Spakovsky allowed that not every improper voter should be prosecuted, but insisted that “each ballot they cast takes away the vote of and dilutes the vote of actual citizens who are voting. And that’s —”

The judge interrupted again. “So, the thousands of actual citizens that should be able to vote but who are not because of the system, because of this law, that’s not diluting the vote and that’s not impairing the integrity of the electoral process, I take it?” she said.

Von Spakovsky didn’t engage with the hypothetical. He simply didn’t believe it was happening. “I don’t believe that this requirement prevents individuals who are eligible to register and vote from doing so.” Later, on the stand, he’d tell Ho he couldn’t think of a single law in the country that he felt negatively impacted anyone’s ability to register or vote.

Robinson, in the end, strongly disagreed. As she wrote in her opinion, “the Court finds that the burden imposed on Kansans by this law outweighs the state’s interest in preventing noncitizen voter fraud, keeping accurate voter rolls, and maintaining confidence in elections. The burden is not just on a ‘few voters,’ but on tens of thousands of voters, many of whom were disenfranchised” by Kobach’s law. The law, she concluded, was a bigger problem than the one it set out to solve, acting as a “deterrent to registration and voting for substantially more eligible Kansans than it has prevented ineligible voters from registering to vote.”