Fact-checking: does anyone even care?
Last year ended with a lot of discussion among media types in the U.S. and abroad about whether fact-checking actually matters. Two things set off the flurry of articles: Donald Trump's brazen rejection of the findings of fact-checkers and The Washington Post's Caitlin Dewey shutting down her “What was fake on the internet this week” feature.
Here are a few headlines representative of this sub-trend in media commentary:
- “The Limits of Fact-Checking: Calling Trump (and others) out for their lies doesn’t seem to make a difference. What’s going on?” (POLITICO Magazine)
- “Why Fact-Checking Donald Trump Backfires: Psychologists explain his enduring support despite his gaffes and lies.” (Slate)
- “Fact-checking, le malaise: Dans une société fracturée, la vérité du fact-checking en est-elle vraiment une?” (Libération)
These are all thoughtful articles that refer back to academic research in constructing their arguments. Moreover, the pieces are overall far more nuanced than their headlines. Nonetheless, they offer incomplete pictures of the effect of fact-checking.
No one will be surprised to read the person who is running the International Fact-Checking Network argue that reports on the imminent demise of fact-checking are overrated. Just as J-Lo seems to believe that she can be as popular an actress as she is a singer, I would be expected to defend fact-checking even if all evidence pointed the other way.
That is not the purpose of my article. There is extensive research on how people deal with mistaken information and the impact of fact-checking. The reality is less gloomy than recent articles portray it even as it is perhaps less hopeful than fact-checkers may wish. It is certainly less black-and-white.
There are two crucial considerations when it comes to the effect of fact-checking on readers. The first is whether fact checks actually reach readers. The second concerns the likelihood that fact checks will convince those who do read them even if it contradicts their previously-held views.
Reaching the reader
Social media echo chambers are a real problem for the circulation of truthful information. A recent study of 54 million U.S. Facebook users showed that "two well-formed and highly segregated communities exist around conspiracy and scientific topics – i.e., users are mainly active in only one category."
And if you think that these communities are split evenly, think again. A 2013 study by researchers at the University of Washington and Northwest University that tracked the diffusion of misinformation on Twitter in the wake of the 2013 Boston Marathon bombings found that the circulation of corrections is "muted compared with the propagation of the misinformation." This means that far more people spread rumors than they did subsequent corrections. In a pithy take, Dewey summed this up in a recent Q&A: "of course the sexy lie is going to get more attention than the very unsexy debunking."
Another obstacle in the attempt to make fact checks reach readers is confirmation bias, which leads people to give disproportionate preference to information that reinforces their beliefs.
Yet while fact checks may not reach everyone, they have been shown to be a format that appeals to readers. A 2015 study found that in the U.S. "[the] public has very positive views of fact-checking and, when randomly exposed to it, comes to view the format even more favorably."
Moreover, 2015 has seen traffic increase enormously at the major U.S. fact-checking operations. Factcheck.org's pageviews in 2015 were up 80 percent compared to 2014 and 227 percent compared to 2011, which is perhaps a better indicator given that it was another non-election year preceding the presidential election. The Washington Post's Fact Checker is reporting similar results, with November 2015 recording three times more pageviews than November 2012 and seven times more than November 2011. Of the 50 most-read Fact Checker posts from the last five years, 31 are from 2015. PolitiFact.com recorded its best year for traffic in 2015 since launching in 2007.
Crucially, these figures don't tell us who is reading the fact checks and to what extent this readership overlaps with those subjected to the original, erroneous, statements. They do however indicate that fact-checking in the age of Trump is increasing its audience, rather than losing readers.
Changing readers' minds
When it reaches readers, is fact-checking convincing them? More specifically, does a reader change his or her opinion if it was proven to be wrong by a fact check? That depends, but research has found that corrections do change people's minds.
This may be less true for the most avid Clinton or Trump supporters faced with a fact check of their political idol: in what has been dubbed the "backfire effect," some corrections can actually reinforce misperceptions.
A recent study struck a more hopeful note about partisanship. The paper, published in December in the Quarterly Journal of Political Science, notes that "the apparent gulf in factual beliefs between members of different parties may be more illusory than real." Researchers found that while factual questions get significantly different responses depending on whether the respondent was a Democrat or a Republican, polarization decreased by a wide margin when there was a small financial reward for giving the correct answer (or replying "don't know"). Partisans seem to be "cheerleading" their side when responding to surveys but are in reality often aware of the facts and their own ignorance.
One thing that does seem to matter is the style used for corrections: Avoiding negations and minimizing the repetition of the false claims are more likely to be effective.
What emerges from research is a relatively complex picture that some of the fretting has been doing disservice to. For as long as we take the presence of false assertions or the success of those who spread them as evidence that fact-checking is done for, we are not concentrating on what really matters. Brendan Nyhan, who has authored a few of the studies quoted in this piece, recently made a less academic but equally effective point:
Saying fact-checking failed b/c pols still lie is silly. Have cops failed b/c crime still exists? Why haven't firefighters eliminated fire??
— Brendan Nyhan (@BrendanNyhan) December 24, 2015
Fact-checkers won't reach or change every mind and a scenario where they did strikes me as not just unrealistic but dystopic. In no human endeavor all that we look at are the facts and subjectivity is banned.
Still, we should be dedicating our best efforts to studying how fact-checking can reach wider audiences and how it can be produced to more effectively correct misperceptions. Dismissing the entire effort as futile because facts keep getting mangled by politicians (and voters keep believing them) makes for powerful headlines but imperfect analysis.