Asking PolitiFact to Police ‘Fake News’ on Facebook is a Terrible Idea

It’s come to this. Facebook has been so bullied over the “fake news” narrative since the election that they’re actively appointing a panel of censors to police speech on Facebook:

To combat fake news, Facebook has partnered with a shortlist of media organizations … that are part of an international fact-checking network led by Poynter, a nonprofit school for journalism located in St. Petersburg, Florida. … Of the 42 news organizations that have committed to Poynter’s fact-checking code of ethics, Facebook is starting out with the following four: Snopes, Factcheck.org, ABC News, and PolitiFact.

In fairness, Facebook says it’s giving these organizations a mandate only to deal with clear “hoaxes,” and they are expressly not to litigate partisan disputes. (Whether or not partisan disputes will get redefined as hoaxes at some point in the future is an obvious concern.) To the uninitiated, these may even seem like reputable news organizations. Conservatives concerned about media bias would likely—and rightly, to some extent—be wary of all of these outlets. But one of these outlets deserves singling out to explain what a horrifying idea this is: Poynter’s very own PolitiFact, which might well be the most biased news organization in America.


The entire media fact-checking enterprise is suspect, and I’ve written at length why this is the case. In sum, it’s basically a way for a bunch of reporters with no particular expertise to render pseudo-scientific judgments on statements from public figures that are obviously argumentative or otherwise unverifiable. Then there’s the matter of them weighing in with thundering certitude—pants on fire!—on complex policy debates they frequently misunderstand.

Even accepting the premise that media “fact-checkers” are bad, PolitiFact has taken a bad idea and driven it to inconceivable levels of awfulness. Let’s just take gander at a few highlights.

When President Obama repeatedly lied about Obamacare enabling people to keep their health insurance, PolitiFact rated it true six different times. The next year, after he was reelected and Obamacare was a rolling disaster, PolitiFact made “if you like your health insurance, you can keep your health insurance” the 2013 lie of the year. They hated being forced to admit such an egregious partisan error, so they bizarrely spent a lot of time chastising Glenn Beck instead of taking responsibility. On nearly all Obamacare-related matters, PolitiFact acted as the propaganda arm of the White House. Here’s the Obama campaign “Truth Team” tweeting out a PolitiFact ruling calling Florida governor Rick Scott a liar for saying that insurance premiums will rise under Obamacare. (Spoiler: They’ve gone up quite a bit.)

Of course, 2013’s lie of the year was a slight improvement over the lie of the year in 2012, which slammed Mitt Romney for a campaign ad about the production of Jeep vehicles in China. After I called them on it, they actually conceded the statement they had chosen as lie of the year was the “literal truth.” In general, their defense of Obama and attacks on Romney were completely credulous. In one notable instance, they refused to consider that the Obama campaign was being intentionally dishonest in their accusations and instead rated their attacks on the Romney campaign’s Medicare plan “half-true” because they were granting the Obama campaign “wiggle room.” Obama really could do no wrong as far as PolitiFact was concerned. As late as 2014, PolitiFact was labelling Obama’s efforts to reform the disastrous Veterans Administration a “promise kept.”

It wasn’t just a matter of defending Obama, either. Downballot, PolitiFact had it out for Ohio’s 2012 Senate candidate, Josh Mandel, giving Mandel three “pants on fire” rulings in the span of a few months. (For comparison’s sake, Harry Reid had been given two “pants on fire” rulings in five years by the organization.) The likely explanation is the extreme bias of Cleveland Plain-Dealer columnist and PolitiFact writer Tom Feran, who authored all three rulings. He had a private Twitter account he used to praise Obama and Occupy Wall Street repeatedly and floridly, call conservatives “wingnuts,” and tweet links to over-the-top, left-wing blog posts on “The Cancer of Conservatism.” After personally giving Mandel three “pants on fire” ratings, the Cleveland Plain-Dealer published yet another article by Feran headlined, “Campaign attacks give Josh Mandel Pants on Fire crown.” When Feran was asked about his Twitter account, he didn’t respond and quietly scrubbed information identifying him as working for the Plain-Dealer from the account.

What’s amazing is just how transparently incompetent PolitiFact rulings often were. They spent years trying and failing to debunk a Mitt Romney talking point about how the Navy was at its smallest size in a century. At one point they made this inane assertion:

And our allies and foes alike have kept their word to the 1922 Washington Naval Conference, which limits the world’s navies by tonnage as a solution to an early arms race. In other words, the Navy has been small since WWI not because of any single administration decision but due to a nearly century-old disarmament accord.

The Washington Naval Conference treaty has been defunct since 1936.

More recently, you can probably guess how they handled 2016. Yes, very badly. PolitiFact rated the claim that Hillary Clinton never sent or received classified information on her private email server “half-true” and later offered some credulous wordsoup that they couldn’t change the half-true rating even after FBI director James Comey confirmed what we knew all along about her sending and receiving classified information. They eventually did reverse themselves, but it was embarrassing even to read the logic PolitiFact was employing. It’s also worth noting that PolitiFact employed several stealth edits to cover up their incompetence.

And, in perhaps the most questionable thing PolitiFact has ever done, PolitiFact did a dubious, error-riddled “fact-check” in September of a Daily Caller News Foundation report on corruption in a Clinton Foundation program in Africa. As it happens, there was “a partnership between PolitiFact and another group ‘to fact-check claims about global health and development’ … funded with $225,000 from the Omidyar Network [the giving arm of eBay founder Pierre Omidyar and his wife]. Pierre Omidyar and his wife, Pamela, gave $1 million to the Clinton Foundation’s HIV/AIDS drug distribution program”—the very same program at the heart of the Daily Caller News Foundation report. There’s no hard evidence PolitiFact was engaged in pay-to-“fact-check” here, but it sure looks bad.

Aside from this parade of journalistic horribles, the irony here is that PolitiFact’s modus operandi—put a statement or fact that’s in the news at the top of the page, followed by a graphic pronouncing it true or false, accompanied by a short explanation—has actually given us a way to verify and quantify media bias. In fact, two university studies have done exactly that. In 2013, the Center for Media and Public Affairs (CMPA) at George Mason University looked at 100 PolitiFact rulings over five months and found the organization rates Republicans liars at a rate of more than two to one over Democrats. The CMPA study also specifically looked at PolitiFact rulings on statements of the 2012 presidential candidates and their surrogates. Again, they were much harder on Republicans:

PolitiFact rated Democratic statements as “mostly true” or “entirely true” about twice as often as Republican statements—42% true ratings for Democrats vs. 20% for Republicans. Conversely, statements by Republicans were rated as entirely false about twice as often as Democratic statements—29% false ratings for GOP statements vs. 15% false ratings for Democrats. (This includes categories labeled “false” and “pants on fire.”)

And, by the way, these results seem consistent with a prior University of Minnesota study:

A Smart Politics content analysis of more than 500 PolitiFact stories from January 2010 through January 2011 finds that current and former Republican officeholders have been assigned substantially harsher grades by the news organization than their Democratic counterparts. In total, 74 of the 98 statements by political figures judged “false” or “pants on fire” over the last 13 months were given to Republicans, or 76 percent, compared to just 22 statements for Democrats (22 percent).

How does PolitiFact explain these discrepancies? Well, there was a weaselly press release when the CMPA study came out. I didn’t think it was terribly satisfactory, so I went to an event at the National Press Club and asked Bill Adair, the former head of PolitiFact, to explain these university studies. He replied that he thinks the organization is fair because he gets complaints about their rulings from both Republicans and Democrats when he goes to parties. No, really. That’s what he said:

You know I don’t find the “hey, you gave my team…” [complaints persuasive] because I hear it from both sides. I was at a party over the summer and a guy came up to me and said, “Hey, I think that, I really think PolitiFact Virginia has been unfair, they’ve been very biased against Tim Kaine,” the Democrat. And then like a week or so later the Virginia Republican party came out and said PolitFact Virginia is unfairly targeting the Republicans.

The charitable interpretation here is that PolitiFact, for whatever reason, just happens to be randomly selecting more false statements from Republicans to fact-check. But a close examination of those individual PolitiFact rulings suggests that the organization is likely biased, dishonest, incompetent, and corrupt. Indeed, we’ve even seen PolitiFact rate the veracity of the exact same statements differently depending on the partisan bent of the politician:


Even thinking of formally censoring things on a wide platform such as Facebook is a terrible idea to begin with, and Facebook has tapped the absolute worst organization I can think of for the job. There’s simply no way PolitiFact—and, by extension, its parent organization, the Poynter Institute—can be trusted to sort out what’s true and what’s not on Facebook.

In theory, maybe media fact-checkers could do some good. But as we’ve seen in years of practice, this entire enterprise is a scam, designed to make it look like if you oppose a demonstrably partisan, self-appointed media referee, you’re against “fact-checking.” If you care about honest discourse, Mark Zuckerberg and Facebook need to hear the message loud and clear: You’re not entitled to your own PolitiFacts.

Related Content