Here’s what I learned about fact-checking elections after embedding with PolitiFact

In an interview to one of Israel’s leading radio stations, the minister for regional cooperation, Tzachi Hanegbi, was asked about Prime Minister Benjamin Netanyahu’s failure to fulfill his 2009 election promise of ending Hamas’ rule in the Gaza Strip.

“What was said at the time is irrelevant,” he said. “In the opposition, you talk, but in government, you act.”

“So, you’re saying that we shouldn’t take what politicians say before elections too seriously?” the interviewer asked.

“Do you think we should? Of course we shouldn’t,” Hanegbi responded.

As fact-checkers, we tend to take the things politicians say very seriously. And while facts always matter, they seem to be even more crucial before elections, when citizens are called to make informed decisions about their elected representatives.

The Whistle, the Israeli fact-checking website where I work as a reporter, is a relatively young organization. First launched in 2017, The Whistle has not yet had a chance to fact-check a campaign. In order to prepare for the upcoming 2019 Israeli general election, we sought to learn from the experience of our fellow U.S. fact-checkers at the (Poynter-owned) PolitiFact.

Seeing as American citizens are called to the polls at the federal level every two years, 10-year-old PolitiFact has had plenty of experience covering campaigns. In the final stretch before the 2018 midterm elections, I visited PolitiFact’s Washington, D.C. bureau as an International Fact-Checking Network fellow. Here’s what I learned in my time with them.

Focus your coverage

With 35 Senate seats and 435 House seats up for grabs, as well as 36 gubernatorial races, PolitiFact’s national team of eleven journalists had to set priorities. They identified 16 Senate races that seemed likely to be especially tight.

“Some of these races turned out to be more interesting than others,” said PolitiFact editor Angie Holan, “and some campaigns were more fact-based than others.”

Once they picked the races they were going to cover, the team started studying the local landscape: who were the candidates, what was the local readership in each state like and what were the hottest topics and main messages of the campaign.

RELATED ARTICLE: There was less misinformation during the midterms than in 2016. But its form has changed.

In some cases, reporters were already familiar with the local issues and politics in the states they covered. In others, they had to start from scratch. Reporter Amy Sherman, who covered Ohio and Florida, said that while she was well acquainted with Florida, having covered the state for several years, Ohio was entirely new to her.

“I had to invest more time, read more stories, pay more attention,” she said.

Reporter John Kruzel, who focused on the Montana Senate race, said his preferred method of getting to know the candidates and their positions was studying their voting records, which he said, “can teach you a lot more than their statements.”

Find your allies

Alongside the Senate races covered by its national team, PolitiFact also covered some House and gubernatorial races, with help from partners throughout the country.

PolitiFact’s local affiliates covered gubernatorial races in Florida, Wisconsin, Texas and California. Some of the closer House races were covered by two freelance journalists. PolitiFact also worked with local “trackers” in congressional districts, who monitored the local races and sent the editors weekly updates on important developments and potential claims to fact-check.

PolitiFact also developed partnerships with West Virginia University and the University of Missouri, where journalism students study and practice fact-checking.

Louis Jacobson, a reporter at PolitiFact, leads the fact-checking course at WVU. The students checked the Senate and congressional district races and other state officials, and he edited their work. The class collectively determined the ratings of each fact check. Since they live in West Virginia, the students are well positioned to hold candidates to a high standard of accountability.

“I tell the students to identify themselves both as PolitiFact journalists and as UWV students,” Jacobson said. “They’re constituents, too, and the candidates should be conscious of that.”

Let checkable statements come to you

Election time is peak season for fact-checkers.

At other times, PolitiFact’s reporters might need to dive into traditional and social media in search of statements to check, but in the period leading up to an election, contentious claims tend to come to them. Campaign ads often contain factual statements, ranging from true to absurd, and facts and figures are constantly thrown around in debates. If the staff missed out on an interesting claim, PolitiFact readers might get in touch with them and suggest they check it.

RELATED ARTICLE: PolitiFact at 11: After 15,000 claims, the truth still matters

Not only are questionable claims found in abundance in the weeks leading up to an election — but politicians and pundits also tend to repeat claims over and over again.

For PolitiFact, this has two main implications. First, although each reporter focused on different races, it was useful to follow everyone’s work and make use of the archive when possible. Second, when a public figure repeated the same claim several times over a short period of time, the journalists at PolitiFact could pick its most straightforward and fact-based formulation.

Keep sight of the big picture

Fact-checking the midterms in different states required PolitiFact’s reporters to acquaint themselves with very specific local issues.

How has Florida Governor Rick Scott’s stance on immigration changed over the years?

How easy is it to vote in North Dakota?

And did driving licensing policy lure illegal immigrants into Tennessee?  

However, over time, it becomes possible to recognize common themes and recurring claims and to gain a broader perspective on the midterms. PolitiFact capitalized on this accumulated knowledge to publish a series of “big” stories on the election.

The project identified the top 10 storylines of the 2018 midterms, drawing attention to the themes they covered most extensively; it wrote an explainer on the top issues at stake, such as healthcare, gun control, trade tariffs and immigration. It dedicated a story to the top false statements made by the President while campaigning for members of his party. It also published some rolling stories, including a guide to the 16 Senate races, and one dedicated to Donald Trump’s campaign rallies, where it responded to each false claim with a short and straightforward comment. Here, too, familiarity with the archive was key, as these responses were often based on existing checks — as was PolitiFact’s live fact-checking of the rallies on Twitter.

Alongside its big stories, PolitiFact also followed and fact-checked the stories dominating national news.

In the weeks before the election, the migrant caravan moving from Central America towards the U.S.-Mexico border was making headlines and generating a swath of speculation and misinformation. Two weeks before the election, 12 explosive devices were sent to prominent critics of the Trump administration. This, too, provided fertile ground for rumors and false news, like the claims that the bombs were a false flag orchestrated by the “Deep State” or that the pipe bombs were never actually mailed.

Although they did not emerge from any of the local campaigns, the false narratives had the potential to influence voting patterns in the different states. Even when focusing on an election, other news events shouldn’t be neglected.

The moment of truth is also the moment of false news

On Election Day itself, PolitiFact shifted its focus somewhat.

According to Holan, by this point politicians had done all their messaging, and were now focusing on getting their voters to the polls. In some of the sketchier corners of the internet, however, fake news was being generated at full speed:

The Black Panthers have taken over Atlanta!

Millions of fraudulent votes have been cast in California!

Stacey Abrams is a communist AND a Muslim Brotherhood sympathizer!

The fact-checking process around false news is slightly different than the one applied to politicians’ statements. Here, fact-checkers may be required to prove that something did not happen, which can be a particularly challenging task. While bad Photoshop jobs can certainly be a giveaway, in order to completely debunk false news stories, fact-checkers often need to go down the rabbit hole in search of their source.

This might entail following headlines and reporting to see how versions have changed over time, in order to figure out what was the basis for the story and how it has been manipulated. It might entail tracking the history of a suspicious picture posted online, which would often turn out to be old photos used in new contexts. And sometimes, it involves even getting in touch with the person who spread the story and asking about their sources – surprisingly, they might themselves admit that their story is groundless.

Most of PolitiFact’s work on this type of false news is done as part of its partnership with Facebook, launched in December 2016. (Disclosure: Being a signatory of the IFCN’s code of principles is a necessary condition for joining the project.)

Holan said she’s optimistic about PolitiFact’s prospects of handling false news in the 2020 presidential election. As the shared effort to reduce the circulation of false news on social media seems to be effective, she said that she would like to increase PolitiFact’s focus on it, and have more staffers focusing on media misinformation. 

Political talking points may change from one election to the next, but fact-checkers like PolitiFact are adapting and learning to cope with new forms of misinformation.