[Note: See the update at the bottom.]
By the very criteria used to proclaim that something is “fake news,” it turns out that the news about “fake news” is…fake news.
That’s the take-away from an investigation of the supposed “fake news” phenomenon – at least as it is being presented in the two canonical posts referenced by every MSM or left-wing outlet that has reported on “fake news” in the last week.
The two posts are from BuzzFeed, and both were penned and/or supervised by Craig Silverman, founder of BuzzFeed Canada, who since July 2016 has been leading a new BuzzFeed initiative to combat “fake news.”
- Hyperpartisan Facebook Pages Are Publishing False And Misleading Information At An Alarming Rate (posted on 20 October 2016)
- This Analysis Shows How Fake Election News Stories Outperformed Real News On Facebook (posted on 16 November 2016)
Three analytical errors make the claims about “fake news” fake.
1. The two posts in question conflate genuinely fake news with articles that draw largely partisan conclusions, about demonstrably valid pieces of news. The same conflation is evident in the many spin-off articles about “fake news” over the past few days, which cite the posts from BuzzFeed (e.g., here and here).
The conclusions BuzzFeed deems “false” or “fake” may sometimes produce misleading headlines. But “misleading” itself, as it pertains to politics, is usually in the eye of the beholder.
The very point of politics is that we don’t all agree on conclusions, and on what is “misleading.” We certainly don’t agree that what some other guy considers “misleading” equals “fake.”
On the other hand, there are legitimate “fake news” sites out there. Such sites and their products should not be conflated with sites that draw politically tendentious inferences from documentable events. (More on that below.)
But that’s exactly what promoters of the “fake news” theme are doing.
2. Conclusions and opinion are labeled “false” in the “fake news” analyses by BuzzFeed. This is our big First Amendment red flag. Conclusions and opinion can’t be false. They can be disputable. Only factual assertions can be labeled false.
But, as we’ll see, in a number of cases, BuzzFeed has to be labeling conclusions and/or opinions false, to designate as “false” articles that contain no factual falsehoods.
3. BuzzFeed mixes apples and oranges by including in their analysis meme images from Facebook that contain no factual content. By definition, an opinion meme isn’t “news,” fake or any other kind. Judging memes by the same standard as news items changes the whole proposition of the analysis, but without notice or explanation.
Of course, polemical writing can’t be judged by the same standard theoretically applied to “straight” news items either – and virtually all of the textual, “information” posts from partisan websites considered in the BuzzFeed analysis are exactly that: polemical writing. They are overtly intended to persuade.
But adding that point together with the dubious inclusion of meme images turns the entire BuzzFeed analysis into an impressionistic exercise in “listing things that annoyed us on Facebook.” It becomes sloppy and inconsistent.
Clearly, BuzzFeed is concerned that a lot of different political items posted to Facebook color users’ perceptions. But in calling many of those items “false” or “fake,” BuzzFeed’s analysts engage in exactly the tendentious eliding of fact and opinion that they purport to be calling out.
A brief survey
For the first piece, from October, BuzzFeed “fact-checked” 2,282 posts put on Facebook from 19-23 and 26-27 September, by nine much-visited Facebook pages: three “mainstream,” three left, and three right. The results can be viewed at the BuzzFeed link above. Each of the 2,282 posts was labeled afterward as “mostly true,” “mostly false,” a “mixture of true and false,” or “no factual content” (these last were the meme images).
In the second piece, from this past week, BuzzFeed looked at the Facebook performance of what BuzzFeed deemed to be the top-performing “fake news” posts about the election on the Facebook site. So this second analysis considered far fewer individual posts, but presumably used the same methods as in the October article to identify them.
The ultimate purpose of the second analysis was to compare the Facebook performance of the so-called “fake news” posts with that of mainstream posts about the election.
An inauspicious beginning
To begin with, it’s important to note that in the October article, BuzzFeed’s own top-featured example of “fake news” is not fake, but at most has one disputable interpretation.
BuzzFeed pointed to a Freedom Daily post that showcased a video of at least one man trying to set two other men on fire. (Ultimately, one man was charged, although it wasn’t clear at the time how many were involved.) The perpetrator was black; the two other men, with skin colors well within the “white” range, Freedom Daily referred to as white. One of Freedom Daily’s points was that the media censored references to the races of the perp and victims (hence the site’s headline allusion to media censorship: “Two White Men Doused With Gasoline, Set On FIRE By Blacks – Media CENSORED”).
BuzzFeed called this post “mostly false.” The reasons: BuzzFeed felt that one of the victims was actually a light-skinned black man; the media had covered the event at the time it happened in 2015; and Freedom Daily mentioned the (documented and linked) endorsement of other, similarly violent attacks by Black Lives Matter protesters. BuzzFeed characterizes that as Freedom Daily saying the perpetrator in the arson video was a BLM member.
BuzzFeed is wrong in what it says about that last point. Freedom Daily did not make that claim.
The disputable interpretation in this case is whether one of the victims was white or black. Freedom Daily documented that the victim’s father is white, and his mother black. Freedom Daily also offers a genetics case, cited and accepted in mainstream sources, for considering the victim white.
The question of how much it even matters certainly occurs to one forcibly – but the point at hand is that this is a disputable matter, not a case of Freedom Daily having said anything false.
Nothing about the Freedom Daily post is demonstrably false. Plenty of people on the right as well as the left would find the tone of the post off-putting. A lot of people would rather focus the public debate elsewhere.
But BuzzFeed’s top example of “fake news” is nothing of the sort. It’s something that actually happened, and the main assertions made by Freedom Daily can be documented and argued for credibly. BuzzFeed doesn’t like the conclusions Freedom Daily draws, to be sure (and is probably offended by how they are presented).
But that’s the whole point of freedom of speech – and it’s exactly the difference between “free speech” and “fake news.”
Another BuzzFeed example is slightly, but only slightly, more problematic. The bottom line about this one is that the headline makes an overly reductionist implication about something that actually happened.
BuzzFeed refers to an article about polls conducted in Australia, in which 49% of respondents said they would favor a ban on Muslim immigration. (Freedom Daily posted this article, from another website, at its Facebook page.) The article’s headline: “Australia Voted to Ban Muslims and Liberals Are Pissed.”
Inelegant, certainly. Not a headline I would use. As BuzzFeed points out, it could be read as suggesting that Australians had voted on a national policy question, and had decided it in some way. Whereas what actually happened was that Australians were polled, and 49% of them favored restricting or banning Muslim immigration.
That is quite clear if you read the post. If the headline is misleading, it’s actually less misleading than many a headline the mainstream media get criticized for (or see here, or do your own searches. Examples are all over the place, and have been for years).
But again, the bottom line is that the headline, which elides two concepts (voting and polling), is arguably too reductionist for what the article actually reports.
If that’s “fake news,” however, then so is BuzzFeed’s reporting on “fake news.” Eliding concepts? Check. Reductionist characterization (i.e., of disputable assertions, based on opinion or analysis, as “fakery”)? Check.
A word on what actual “fake news” is
News that’s actually fake is news that’s made up. Fabricated. Never happened. Concocted out of thin air.
Some of it is perpetrated by pernicious jokesters, like the guy allegedly named Paul Horner who has hoodwinked a lot of people with his faked report about a Craigslist ad for rent-a-mob protesters. Or like the people behind the “satirical” website National Report.
Some fake news is in the well-known “Pallywood” category: created with staged video and photos to foster a false narrative. Actual people may have been present and captured in imagery, but what they’re purportedly seen doing is a false tale. Alternatively, images are photoshopped, or used out of context (e.g., when images from the civil war in Syria were posted as if they were from Gaza, in the summer of 2014).
A key characteristic of fake news is that its falsity can be reasonably verified. When the “dead body” from a Pallywood video keeps wriggling out of its blanket and looking very much alive, you’ve got yourself a case of fake news. When a National Report story says Obama kept a Muslim museum in Washington, D.C. open during the government “shut down” of 2013, while other offices were closed down, you can verify pretty easily that no such thing happened.
A more problematic form of fake news comes from coyly anonymous websites that claim to have inside information about world events. The notorious Debka-file website is one of the best known; another is WhatDoesItMean.com, which features the pseudonymous “Sorcha Faal.” These sites usually deploy enough verifiable information from mainstream sources to invite credulity, but then make “insider” assertions that can’t possibly be tested and disproven, because no one actually has that kind of knowledge.
There is, in fact, a lot of fake news out there. But to lump in with such sites everyone whose opinions you dispute is to be guilty of the very thing BuzzFeed decries. The more you disagree with something, the more simplistic, disingenuous, or manipulative its presentation may seem to you. But it’s not “fake” just because you don’t like the conclusions someone else is using rhetorical tools to showcase.
Some further counter-analysis
To investigate BuzzFeed’s analysis and conclusions, I delved into the October article, and took a sample of the first 10 items from Eagle Rising that were labeled either “mostly false” or a “mixture of true and false” in the BuzzFeed spreadsheet.
What was it that caused the BuzzFeed team to label these posts as it did? In each case, it was not obvious that the post contained actual falsehoods. Instead, the posts in question contained interpretations, conclusions, or (in one or perhaps two cases) assertions of fact that can be considered disputable. People with different political opinions will differ on how disputable they are. But in no case – not even the post about Obama’s birth certificate – is an assertion made that is demonstrably false, by an indisputable standard of evidence.
So that you can replicate the same investigation, I include the screen cap with the posts I looked into numbered and highlighted.
We’ll consider three of the items here. Each of the ten had similar features: factual assertions documented with links and/or video, and interpretations or conclusions that BuzzFeed apparently disagreed with.*
Three sample cases
Number 1, labeled “mostly false” by BuzzFeed, is a Proud Con post about Hillary Clinton’s campaign enlisting illegal “Dreamers” to assist in a voter registration drive. The headline reads “Fox News Just Exposed Hillary’s Illegal Voting Scheme That Rigs Election Against Trump.”
The main claim of fact, that the Hillary campaign was using Dreamers in a voter registration drive, was reported by Fox News. The Fox report quoted from the Clinton campaign’s own literature and used reporting from AP as well. The Dreamer-involved voter registration drive was not in question.
Proud Con said that in doing this, the Clinton campaign was violating U.S. law by harboring illegal aliens. Proud Con implied the same thing in pointing out that Clinton had illegals speak at the Democratic Convention in July.
The interpretation of law, and whether illegal aliens were being “harbored,” may be a disputable matter. But this claim is not manifestly false. BuzzFeed may simply have a different political perspective on it.
Proud Con’s headline also took a conclusion as its premise: that in using Dreamers to get out the vote, Hillary was trying to affect the vote unlawfully (“rig the election”) by having ineligible illegals cast ballots.
That is not manifestly false. It is disputable. If a presidential candidate enlists people who are in the country illegally to register voters, the question why is a perfectly valid one. And there’s nothing unreasonable about supposing that the intended voter-registration outreach is to people who are also in the country illegally. There is not, nor can there be, evidence that biased assumptions about those questions are wrong.
There’s nothing “mostly false” about the Proud Cons post. There are only interpretations and conclusions that BuzzFeed disagrees with.
If that were fake news, the New York Times, CNN, and Politico would be guilty of it every day. (E.g., when they report religious-liberty pushes prejudicially as “anti-gay activism.”)
It’s not fake news. It’s free speech.
Case number 4 is a post by The Politistick about a surge in Donald Trump’s poll numbers in mid-September. BuzzFeed calls this one a “mixture of true and false.” The headline reads: “SHOCKING LA Times Poll Shows Trump Leading Big, Support Soaring Among Black Voters.”
Politistick, writing in mid-September, says the poll showed Trump surging to a big lead after Hillary nearly collapsed at a 9/11 event. The poll presentation now shows the historical results all the way through Election Day, but the surge described by Politistick in September is easily discernible. That assertion checks out.
Politistick’s claim that support for Trump increased at the same time among black voters also checks out.
I’m guessing it’s this next part that BuzzFeed deemed “false.” The post alludes to Hillary “going full force with the ‘birther equals racist’ narrative,” and attributes that to Democratic worries over the increase in black support for Trump.
Hillary was, in fact, personally emphasizing at the time a narrative that it was racist to question Obama’s place of birth. So that checks out too.
It is an interpretation of facts, on the other hand, for Politistick to suggest that Hillary’s campaign was flogging this theme because of worries about the black vote shifting to Trump.
An interpretation is analytically disputable. It’s not “fact-checkable.”
An interpretation also doesn’t have to be heralded with trumpets. Highly regarded pundits from Walter Russell Mead to Charles Krauthammer to Jeffrey Goldberg present interpretations every time they write something, without framing them in asterisks and warning labels. Readers are expected to understand when they are being presented with an interpretation, as opposed to a bald fact. That’s what all those exercises in reading comprehension were for, back in fourth grade.
Interpretations aren’t inherently conspiracies against truth (although at times they may be), nor are they fake news. The ability to formulate them is a tremendous asset to the human race, on balance. They are at the core of our commitment to free speech, even when they aren’t elegant or admirable.
The third case is number 9, an Eagle Rising post (written by Tim Brown of Freedom Outpost) about ICE misrepresenting how many illegal aliens were convicted of “deportable” crimes, but not actually deported.
The headline reads: “Obama Administration “Grossly Misrepresenting” Crimes Committed by Illegal Aliens!” BuzzFeed labels it a “mixture of true and false.” But there is nothing apparently untrue in the post.
Brown quotes a report from the Federation for American Immigration Reform (FAIR), to make the case that ICE misrepresented the numbers in question in public testimony to Congress. A FOIA request made on FAIR’s behalf got federal documents released which showed that the number of “deportable” crimes committed by released illegals in the year 2014 was nearly ten times the figure ICE had briefed to Congress.
FAIR provided documentation of the crimes for which the illegals were convicted, demonstrating that the assessment of the numbers was valid.
The House Judiciary Committee agreed with FAIR’s assessment, in a news release a few days after the FAIR report. The factual claims about the numbers, where they came from, and how they differed all check out.
I’m guessing here, but I suspect that BuzzFeed objected to the following statement made by Brown early in the post:
Now, if the term “illegal, criminal aliens” sounds like overkill, understand that the term illegal does refer to aliens who are in the United States illegally, which makes them criminals. However, applying the term criminal relates to the fact that they committed a crime while here illegals [sic] and were incarcerated, but not deported.
BuzzFeed would probably point out that being present illegally in the U.S. is a civil offense, not a criminal one, and therefore Brown has made an untrue statement here. (You can decide how significant you think this is, since it is not germane to the central thesis of the post.)
But it must also be pointed out that unlawful entry into the U.S. is a criminal offense. Brown may not have made the distinction in his comment, but he is still actually correct about many – probably almost all – of the illegal aliens in question.
You are at liberty, of course, to dispute my conclusion about that point. But Brown’s post can in no way be characterized as “fake news.” Nor is it just or reasonable to dismiss it as a “mixture of true and false.” At the absolute most, it’s a post full of well-documented information with a tangential statement in it that could have been worded more carefully.
I actually expected to find more that was legitimately “false” in the ten posts I investigated. But I didn’t. I found facts that can’t be in legitimate dispute – alongside conclusions, interpretation, and opinion that can.
That finding is simply not compatible with the conclusion that these posts constitute “fake news” flooding Facebook.
In its 16 November article, BuzzFeed did identify a “right-wing” post that I would unreservedly agree is “fake news,” and that’s the one put up by the website EndingtheFed about the Pope supposedly endorsing Trump. But it borders on defamatory to suggest that a post like Tim Brown’s is the same kind of post.
And it crosses over into sloppy, meaningless analysis to conflate Brown’s fact-filled piece on criminal convictions for illegal aliens with opinion memes, and call the whole mixed bag a “fake news” trend.
The two points to reiterate in closing are these. BuzzFeed, in making its case about “fake news,” has done exactly what it claims the “fake news” stories have done.
But BuzzFeed should remain free to do that. It’s free speech – and it’s on the rest of us to make the right distinctions about what other people are saying when they engage in it. Other people’s speech doesn’t need to be reined in or mischaracterized, just because we disagree with it.
When it comes to free speech, the whole point is that opinions, interpretations, and conclusions are not a bug. They’re a feature.
UPDATE: As this goes to post, I see that National Review has published a post entitled “Report on Scourge of Fake News Turns out to be Faked,” by Paul Crookston Although I haven’t read it yet, and performed my investigation independently, I wanted to acknowledge that the NRO post is out there. I doubt that I’d be making a mistake to commend it to your perusal.
* Case number 8 is a post about Obama’s birth certificate, treating specifically the question of whether it is racist to question its validity. BuzzFeed labels it “mostly false.” The post doesn’t have links to back up its undoubtedly biased claims about the birth certificate or other facts from Obama’s life. That said, to call it “fake news” is to assume what has never been conclusively proven: that none of those claims is true.
Of equal importance, the main point of the post is to express the opinion that questioning the birth certificate is not a “racist” thing to do. By definition, that cannot qualify as “fake news.” It’s an expression of opinion, disputable but in no way illegitimate.
One could characterize such a post about the Obama birth certificate in any number of negative ways – tired, overworked, unproven, biased, silly, pointless – but the ones BuzzFeed wants to use are precisely the ones that don’t fit. The post is neither “fake” nor demonstrably “false.” That’s the basic problem with each of the posts I looked into.