Fact-checking under President Trump
Media self-loathing is flooding the internet.
Finger-wagging critiques have been directed at many segments of the industry. Poll aggregators failed to properly represent the uncertainty of the result. Cable news polluted the airwaves with uncritical and senseless Trump coverage. Pundits didn't bother to dig beyond the smart take on their Twitter feed.
Critics will no doubt soon turn their attention to fact-checkers. Having spent a summer flogging the meaningless "post-truth" trope (see GIF below), angsty liberal commentators will embark once again on their nihilistic dismissal of fact-checking.
[caption id="attachment_438631" align="aligncenter" width="525"]"I have a good idea, why don't we call it a 'post-truth' era?" *high-fives self*[/caption]
First things first. Reflecting on the lessons from this campaign for political reporting, Ann Telnaes, The Washington Post editorial cartoonist told Poynter that in the early stages, too many journalists just "giggled their way through Trump interviews."
Fact-checkers cannot be accused of not taking Trump seriously from the get-go. His claims earned early and critical attention from the "Big Three" fact-checkers (Factcheck.org, PolitiFact and The Washington Post's Fact Checker).
But that isn't enough for post-truthers, because Trump won despite all this fact-checking. How is it possible, the canned argument goes, that a candidate so universally billed as untruthful by the fact-checkers could win?
The argument is fatally flawed: Political actors, not the media, are supposed to defeat other political actors. Outsourcing electoral success to fact-checkers is to rob voters of their agency and severely misunderstand the role of the media. Had Clinton won, would it have been fact-checking "Wot Won It?" Of course not.
Still, this has been an election where fact-checking has repeatedly been front and center, especially in the run-up to the presidential debates. Several media experts noted to Poynter that this has been a game-changing election for the media's role as fact-checkers. There are lessons to be learned from how this effort played out, even if they will take more than 24 hours to be fully evident.
Operating in a low-trust environment
It bears repeating that this was a particularly challenging election for political fact-checkers, due to the very nature of the two candidates. Donald Trump and Hillary Clinton were respectively the most and second-most disliked candidates in at least 36 years, per FiveThirtyEight.
When so many voters hold strongly unfavorable views of one (or both) candidates, it is inevitable that their appreciation of fact-checking will be colored by their personal feelings.
[caption id="attachment_438581" align="aligncenter" width="526"] Source: FiveThirtyEight[/caption]
PolitiFact editor Angie Holan encapsulated how this broad unfavorability translated into the public's perception of the candidates' truthfulness back in May.
One candidate is so calculated in how she parses facts, people see her at best as secretive and at worst as a liar. The other candidate is so careless with facts, people see him as at best an entertainer and at worst as a liar.
While we know that fact-checking changes readers' minds, we also know that humans are prone to the confirmation bias (looking for information that supports our views) and motivated reasoning (explaining away information that doesn't). With so many people having such strong negative feelings about these candidates, these unsavory psychological traits kick in.
Fact-checkers are also operating in a context of declining trust in the media. Unlike in the rest of the world, most fact-checking in America is conducted by journalists. The declining tide risks sinking all ships.
[caption id="attachment_438583" align="aligncenter" width="564"] Source: Gallup[/caption]
So much fact-checking
This could credibly be billed the most fact-checked U.S. election ever. The Duke Reporters' Lab found 52 fact-checking initiatives operating across the country this election season. More fact-checking was consumed by more readers: The "Big Three" all broke traffic records this year. PolitiFact reached its 100 millionth page view this year on Election Day. NPR's live fact-checking of the debate transcript was the most popular online offering in its entire history.
Moreover, in polls, Americans generally declared themselves favorable to more fact-checking. Sixty percent wanted moderators to fact-check the candidates on stage. Eighty-three percent of registered voters told Pew that it is the media’s role to fact-check political candidates and campaigns.
If 2016 "made fact-checking great again," it wasn't through volume alone; the field also saw some signs of lasting innovation.
The Duke Reporters' Lab tested a pop-up fact-checking widget over the livestream of the final debate; NPR's annotation tool was so successful because it was a better product than previous efforts; The New York Times called a lie a lie on its front page, exciting some; CNN deployed its sometimes misleading chyron to fact-check Trump.
Still, all this fact-checking did not stop a candidate who belched whoppers about unemployment, Muslim-Americans celebrating the fall of Twin Towers and Barack Obama's birth certificate. What happened?
[caption id="attachment_434943" align="aligncenter" width="3312"] Screenshot of the NBC livestream of the third presidential debate and the FactPopUp widget developed by Duke Uni. student Gautam Hathi[/caption]
Ultra-literal fact-checking and what Trump "really meant"
The main problem with fact-checking this election season is that not all of it has been up to scratch. Trump's most preposterous claims were such easy targets that everyone joined in on the fun, sometimes resulting in fact checks like the one below by NBC News.
This was by no means the only ultra-literal fact check, but it serves as a perfect example. Who on Earth would seriously have thought Trump meant Clinton had physically poured bleach all over her emails? And even if someone did, was this a topic that warrants fact-checking?
The subgenre of fact-checking that tallies the misstatements made by Trump over a specific period of time has been particularly prone to this vice. When Trump said "they've shut Christianity down" and Clinton has "no plan" for the economy, how many of his supporters were actually taking it to mean exactly that? Why did journalists?
The indefatigable Margaret Sullivan wrote in the small hours of Wednesday morning that as unpleasant as it may be, journalists should reflect on Peter Thiel's words about Donald Trump (likely inspired by this Atlantic piece).
The media is always taking Trump literally. It never takes him seriously, but it always takes him literally.
Silly fact checks may in turn have given Trump supporters a more legitimate reason to dismiss the entire effort as biased.
"Fact-checkers' reputation is only as strong as the weakest fact check that a skeptic has come across," says Dartmouth professor Brendan Nyhan.
Fact-checkers as an advocate, not an adversary
Politicians have been 'weaponizing' fact checks long before 2016, but Trump's frontal attacks and communication style drew fact-checking into playing a more adversarial role.
The Washington Post Fact Checker Glenn Kessler disputes this conclusion. "The reason why there were so many fact checks of Trump was that he spoke more [...] We would have liked to publish a lot more fact checks of Hillary than we did but she didn't give many interviews, her speeches were rigidly vetted and didn't vary that much. Meanwhile Trump would call in to 4 to 5 TV shows and go off the script in rallies."
This disparity exacerbated the statistical insignificance of the tallies of both sides' lies. And yet media reports about Trump's problem with the facts often focussed on these tallies.
America is not the only country in the midst of an existential breakdown over the role of facts in public debate. Brexiting Great Britain has also seen its fair share of fact-bashing and "post-truth" headlines.
Full Fact Director Will Moy says his organization has always tried to "play the ball and not the man." While this doesn't seem to have helped with public understanding of facts about the EU ahead of the Brexit referendum, it has mostly isolated them from political tugs of war.
In the long run, Full Fact would like to become more of a "consumer advocate for politics," says Moy. Consumer groups tend to enjoy higher levels of trust because they take citizens' side on issues that concern them directly. Fact-checkers' mission to sort fact from fiction is at heart a similar one to that of these groups, but the effects of a flawed political claim are less directly tangible than those of a flawed washing machine.
The closer fact-checkers are able to tie their work to everyday experience the better, because that's often how political misinformation spreads. Kessler thinks Trump's use of specific anecdotes of undocumented immigrants committing criminal acts often overrode, in the eyes of rally-goers, the need for his overall statistical claims about immigration to be accurate.
"People need to see something of their experience of the world reflected in politics as well as data," Full Fact head of communications and impact Phoebe Arnold. A manual worker in a northern industrial town won't believe that the average effect of immigration on wages is zero to slightly positive: he thinks the figures have been fiddled. And his reaction might have a grain of truth in it: EU immigration has a stronger push-down effect on the wages of people who earn less. "Averages and aggregates aren’t a good tactic for rebuilding trust with the electorate," adds Arnold
"Post-truth" is nonsense, but the reach and impact of fact-checking has to be studied further by practitioners and political scientists.
A recent study conducted at key moments of the campaign concluded that "by and large, citizens heed factual information, even when such information challenges their partisan and ideological commitments." That is heartening, but we need to know more that can be applicable in situations like the campaign that just ended.
Did fact-checking not turn voters away from Trump because they were too safely cocooned in their echo chambers to even read it? If this were the case, working with Facebook and Google will become increasingly fundamental for fact-checkers to maintain their relevance.
Or were voters reached, corrected in their views, but determined to vote for Trump for other reasons? In that case, it is a problem less for fact-checkers and more for political activists.
A cartoon published this summer noted that a Trump presidency would likely create lots of jobs – most of them in fact-checking. Beyond the gag, though, Trump's presidency may see an evolution in his relationship with fact-checkers.
The same candidate that vilified fact-checking as "scum" at a rally was happy to call himself "honored" during a debate because fact-checkers found his claim to be right. That suggests that even Trump finds uttering fact-checked statements politically advantageous when he's in front of an audience not composed solely of his own supporters.
From their side, Holan and Kessler don't think they will be changing the way they fact-check, even though they acknowledge it is too early to draw sweeping lessons from the campaign.
"We will be fact-checking the Trump presidency like we did the Obama presidency: with an emphasis on newsworthy claims and a research method based on primary evidence," says Holan. (PolitiFact had also started revamping its Promise Meter before the Election was over.)
Holan dismissed the sense that fact-checking needs to have electoral consequences to matter.
"As a journalist I think there's intrinsic worth in truth-telling. We have to say what is accurate and what is not, no matter how elections turn out, no matter what happens in the political sphere. I think there is a lot of worth in the simple act of truth-telling."
Correction: an earlier version of this post erroneously indicated fact checks have been "weaponized" long before "2106." While that is technically true, we have not been to the future yet and that should have read 2016. We regret the error.