Fact-checking as such should be very narrow in scope. A false claim about some material fact should be corrected. A number or statistic or historical claim that is easily evaluated to an unambiguous answer is corrected, with some minimal commentary on how the claimant's conclusion changes. If I say there are 200 million people in the United States, such that each person on average owes $100,000 of the $20 trillion federal debt, it is the proper place of a fact-checker to say that the true figure is 323 million people and that the per-person average is closer to $60,000. On the other hand, if I want to argue that the true figure is not $20 trillion (unpaid bills that we have covered through borrowing) but more like $200 trillion (total unfunded liabilities for Social Security, Medicare, Medicaid, and other entitlements), that is not a piece of fact-checking. That is editorializing. It moves beyond the narrow scope of correcting a singular piece of information to telling an entirely new narrative. "The debt" is a fairly well-defined concept. "Total unfunded future liabilities" is another concept. It is arguably more relevant and less arbitrarily defined than "the debt." But changing the subject is very different from fact-checking. Media personalities need to be much more aware of when they are writing an opinion piece.
Editorializing is important, and I'm not knocking it. Some narratives really are better than others. We need journalists and academics to write opinion pieces sorting this out. But an entire narrative cannot be "fact-checked". A narrative can be wrong even though each material piece of data mustered in its defense proves correct. A narrative can be right even though transparent falsehoods are used to support it. Some things can be known with near certainty through sheer logic and introspection. It is impossible to fact-check a statement such as "demand curves slope downward." Getting at the truth requires more than assembling facts, because facts don't speak for themselves. Some assembly is required. Facts must be arranged into coherent stories, and this always requires a hefty dose of theory. There is a strong desire to have "the facts" on one's side. But facts themselves are neutral. They are often consistent with any of several competing, mutually exclusive hypotheses. We should stop pretending that if we simply collect enough of these crystallized nuggets of unassailable truth, we win the argument.
I should probably list a few example of bad fact-checking.
See this Slate story about motorists in China killing accident victims and this Snopes piece "debunking" the Slate piece, plus this China Insider story arguing that the original Slate piece is probably true.
See this story, in which the Washington Post "fact checks" a claim about declining union membership and union support by the labor force; Mark Hemingway of the Weekly Standard calls them out on their shenanigans. This is a great example of what I'm talking about. It is perfectly legitimate to discuss various pieces of evidence and argue about the implications of various historical facts and statistics, but that is emphatically not fact-checking. That is editorializing.
See this Marginal Revolutions piece about fact-checkers disagreeing with each other at an unacceptably high rate (with a link to the academic paper). And this Weekly Standard piece discussing it (also by Mark Hemingway).
Read a few of these posts from the WUWT blog about Wikipedia expelling contrary viewpoints on climate science. Wikipedia is often cited as the ultimate arbiter of truth. Plainly it's not if a few zealots are systematically deleting everything that casts doubt on their worldview. In a certain online forum that I frequent, people routinely fact-check by citing Wikipedia. That's often a good start, but it's never quite safe.
Econlog has some good pieces on political bias in fact-checking. See this one by Bryan Caplan and this one by David Henderson. I like the point Henderson makes in his piece. Sometimes fact-checking is not the journalist's job. For example "correcting" one of the candidates during a presidential debate, while hypocritically getting the facts wrong.
A correspondent on NPR's Morning Edition does this kind of thing all the time. He will instant-fact-check his conservative guests: "Hold on, I'm gonna have to fact-check you there..." And then he interjects with his opinion. His left-leaning guests don't get a similar treatment, even though they are just as likely to spew falsehoods. I want to say, "Buddy, it's not your job to fact-check based on your own fuzzy memory of what the 'facts' are."
Editorializing is important, and I'm not knocking it. Some narratives really are better than others. We need journalists and academics to write opinion pieces sorting this out. But an entire narrative cannot be "fact-checked". A narrative can be wrong even though each material piece of data mustered in its defense proves correct. A narrative can be right even though transparent falsehoods are used to support it. Some things can be known with near certainty through sheer logic and introspection. It is impossible to fact-check a statement such as "demand curves slope downward." Getting at the truth requires more than assembling facts, because facts don't speak for themselves. Some assembly is required. Facts must be arranged into coherent stories, and this always requires a hefty dose of theory. There is a strong desire to have "the facts" on one's side. But facts themselves are neutral. They are often consistent with any of several competing, mutually exclusive hypotheses. We should stop pretending that if we simply collect enough of these crystallized nuggets of unassailable truth, we win the argument.
I should probably list a few example of bad fact-checking.
See this Slate story about motorists in China killing accident victims and this Snopes piece "debunking" the Slate piece, plus this China Insider story arguing that the original Slate piece is probably true.
See this story, in which the Washington Post "fact checks" a claim about declining union membership and union support by the labor force; Mark Hemingway of the Weekly Standard calls them out on their shenanigans. This is a great example of what I'm talking about. It is perfectly legitimate to discuss various pieces of evidence and argue about the implications of various historical facts and statistics, but that is emphatically not fact-checking. That is editorializing.
See this Marginal Revolutions piece about fact-checkers disagreeing with each other at an unacceptably high rate (with a link to the academic paper). And this Weekly Standard piece discussing it (also by Mark Hemingway).
Read a few of these posts from the WUWT blog about Wikipedia expelling contrary viewpoints on climate science. Wikipedia is often cited as the ultimate arbiter of truth. Plainly it's not if a few zealots are systematically deleting everything that casts doubt on their worldview. In a certain online forum that I frequent, people routinely fact-check by citing Wikipedia. That's often a good start, but it's never quite safe.
Econlog has some good pieces on political bias in fact-checking. See this one by Bryan Caplan and this one by David Henderson. I like the point Henderson makes in his piece. Sometimes fact-checking is not the journalist's job. For example "correcting" one of the candidates during a presidential debate, while hypocritically getting the facts wrong.
A correspondent on NPR's Morning Edition does this kind of thing all the time. He will instant-fact-check his conservative guests: "Hold on, I'm gonna have to fact-check you there..." And then he interjects with his opinion. His left-leaning guests don't get a similar treatment, even though they are just as likely to spew falsehoods. I want to say, "Buddy, it's not your job to fact-check based on your own fuzzy memory of what the 'facts' are."
No comments:
Post a Comment