On Monday evening, just as the Iowa caucuses were heating up, the Washington Post published a story with this unambiguous headline: “Conservatives spread false claims on Twitter about electoral fraud as Iowans prepare to caucus.” The story was damning in tone and unequivocal in its assertions. “The claims of electoral fraud were false, proved untrue by public data and the state’s top election official,” it began. “That didn’t stop them from going viral, as right-wing activists took to Twitter over the weekend to spread specious allegations of malfeasance on the eve of Iowa’s first-in-the-nation caucuses.”

While concern about “fake news” influencing elections is a legitimate concern, in its rush to debunk a false claim going viral, the Post itself may be spreading fake news. Even more worrisome, the Post’s bad reporting was used to scrub information from social media.

What happened was this: On Monday, the conservative group Judicial Watch put out a press release headlined, “Eight Iowa Counties Have Total Registration Rates Larger than Eligible Voter Population – at Least 18,658 Extra Names on Iowa Voting Rolls.” This is a pretty standard public relations tactic. Judicial Watch was seizing on the news of the day to highlight an issue it had been working on long before the Iowa caucuses. Last year, a Judicial Watch lawsuit forced Los Angeles County to settle out of court and clean up its voter rolls – the county had a staggering 1.5 million more voter registrations on file than voting-age citizens who lived there. Another lawsuit brought by the group forced Kentucky to remove an excess of 250,000 voter registrations from the rolls, as mandated by federal law.

Conservative activists quickly started spreading Judicial Watch’s claim about Iowa on social media, and Post reporters sprang into action trying to verify it. They contacted Iowa’s Republican secretary of state who disputed the claim by pointing to updated voter registration data on the state’s website. With this data in hand, the Post went to work. “Of the eight Iowa counties listed by Judicial Watch, a single one — Lyon County — has more registered voters (8,490) than adult residents (8,430), based on five-year estimates released by the Census Bureau in 2018,” the newspaper reported.

Verifying the Post’s own reporting here is difficult for a number of reasons. The Iowa secretary of state’s own press release doesn’t say how many counties have more registrations than voters. “Iowa’s voter registration statistics are publicly available on the Secretary of State’s website. They are updated monthly,” reads the release. “These numbers show that the ones claimed by Judicial Watch in their news release today are patently false.” The Iowa secretary of state’s office declined requests to comment further.  

Worse, the Washington Post’s own reporting here is contradicted by … the Washington Post. The paper’s blogger Phillip Bump also jumped on the story Monday and came back with different numbers than either Iowa or Judicial Watch. “There are, however, five counties (Dallas, Dickinson, Johnson, Lyon and Scott) where there are more people registered to vote than there were voting-age citizens counted in the vintage 2018 county population data,” he wrote.

So what is the number of counties with voter registrations exceeding 100% – one, eight, or five? Bump didn’t say, although he sided with Iowa Secretary of State Paul D. Pate over Thomas J. Fitton, president of Judicial Watch.  “Fitton also responded to Pate, insisting that his organization’s data is accurate,” Bump wrote. “It isn’t.”

But it’s not that clear-cut. In reality, the numbers are in dispute because Judicial Watch is using a different methodology for analyzing them. Judicial Watch arrived at its figures using the latest data from the federal Electoral Assistance Commission, which came out last July. The group then hired professional demographers to interpret that data. While the Iowa secretary of state has slightly more recent monthly data for voter registrations, the Post compares these registration numbers against census data from 2018, which is a simple and less sophisticated analysis than would be done by a demographer.

One informed source suggests that perhaps one of the reasons there is a discrepancy in the findings is that Judicial Watch may not have accounted for the for the fact Iowa allows 17 years-olds who will be old enough to vote on Election Day to pre-register -- and there may be as many as 5,000 17 year-olds in the state registered to vote. But Judicial Watch responds that Iowa is one of 21 states that allow pre-registration, and this wouldn't affect the large-scale voting analyses it does. Regardless, Judicial Watch is claiming an excess of 18,000 registered voters in eight countries and 5,000 votes would not explain the difference between eight counties with too many registrations and just one, as the Post contends.

For now, Judicial Watch is standing by its figures. And given the different methodologies involved, a more sophisticated look at the numbers would be required to declare whose numbers are wrong and whose are right. It’s also worth noting both Kentucky and California didn’t contest the accuracy of Judicial Watch’s voter data in federal lawsuits brought against them last year.

That’s in part because conservative groups aren’t the only ones worried about bloated and inaccurate voter rolls. Census data shows more than 10% of Americans move every year, and states and localities don’t do a good job of keeping registrations up-to-date – a 2012 Pew study found that 24 million voter registrations, about one out of every eight, are “no longer valid or are significantly inaccurate.”

Further, the Post’s reporting betrays an obvious political slant. “Flaws in the data did not stop other conservative activists from pushing the misleading conclusion,” it declared. “Charlie Kirk, the founder of Turning Point USA, a group mobilizing young conservatives, followed up Sunday afternoon to proclaim, ‘One day before the Iowa Caucus, it’s been revealed that EIGHT Iowa counties have more adults registered to vote than voting-aged adults living there.’ He asked users to retweet to show their support for a national voter-identification law.”

It simply hasn’t been definitively established that the claim is misleading. And since when is it sinister for activists to draw attention to evidence that supports their side of a political argument? Even if only one county in Iowa had more registrations on file than voters, as the same Post story reports, that could fairly be used as an argument for voter ID laws – which Iowa’s secretary of state supports.

Additionally, the Post mischaracterized what Judicial Watch’s claims about Iowa.  The Post asserted flatly that Judicial Watch had made false “claims of electoral fraud.” While it’s true that having excess and inaccurate voter registrations on file makes states and localities more vulnerable to fraud, Judicial Watch didn’t allege that it was occurring. It raised the specter that Iowa might be susceptible to manipulation, as are all states with inaccurate voting rolls. The aforementioned Pew study found 2.75 million Americans were registered to vote in more than one state, which does lead to some amount of “double voting” – late last year a man was convicted of double voting in Florida and New Hampshire.

The Post was warned about this distinction, but ignored it. “But please be sure you are accurate and stop falsely reporting that we have alleged voter fraud, and put out false information etc.,” Fitton asked Post reporter Isaac Stanley-Becker in an email exchange shown to RealClearPolitics. “Please share my concerns with your editor and, in the least, fix your headline.” Stanley-Becker responded with a one sentence email saying, “My editor says dirty voter rolls are a form of electoral fraud.” (This is in reference to Judicial Watch calling inaccurate voter rolls “dirty.”) Throughout the exchange Fitton repeats that Judicial Watch has not claimed specific instances of voter fraud have occurred in Iowa.

In another instance, Stanley-Becker tries to rebut Fitton by highlighting a press release from Judicial Watch noting that excess voter registrations are a “red flag for voter fraud,” which, as a semantic matter, is accurate and does not assert voter fraud is definitively occurring, unlike how the Post is characterizing Judicial Watch’s work. Stanley-Becker did not respond to a request to comment on the criticisms of his story.   

The effects of the Post’s reporting were not negligible. As a result of its story, Judicial Watch’s posts on Facebook and Instagram were taken down, raising questions about the power of the media and partisan officials to stop certain stories from being circulated – regardless of whether they are accurate or fair -- in their attempts to police facts.

Fitton isn’t hiding his frustration. “The data they're pointing to is actually consistent with our concerns and Facebook deletes our posts and suppresses our Instagram? Because a government official disputes our numbers?” he says, adding, “It's a terrible example of the government pressuring private so-called independent media to suppress something a government official doesn't like.”

Whether Fitton is right about the Iowa secretary of state’s intentions remains to be seen, and the accuracy of everybody’s numbers is still in dispute. What is indisputable is that inaccurate voter rolls are a significant problem in Iowa and the rest of the country. The Washington Post’s reporting seems designed to gloss over the problem rather than shine a light on it.