Tuesday, March 8, 2016


Fact-Checking-Checking Dept. - Politifact is at it again. Earlier today, the fact-checking site decided to examine a comment made by Democratic presidential candidate Bernie Sanders:

"'I would love to run against Donald Trump, and I’ll tell you why,' said Bernie Sanders. 'For a start, … not all, but almost every poll has shown that Sanders vs. Trump does a lot better than Clinton vs. Trump. … And, that’s true nationally.'"

PolitiFact examined Sanders' characterization of the polling data--the only factual assertion on the table--and verified that it was entirely accurate, then ranked his comment as only "mostly true," manufacturing a completely ridiculous rationale for denying Sanders a "true" rating. "[P]olling experts say such results should be taken with a grain of salt, since polls taken well before the start of the general-election contest have historically not been very accurate predictors of the November results." So what Sanders said--in a comment aimed at expressing his desire to run against Trump, not at sizing up the ins and outs of polling--is less than entirely true based on something Sanders never said.

Politifact's writers make a regular habit of this sort of thing. If, as with Sanders, they can't find anything inaccurate about a statement, they'll often just throw it out, invent a straw man claim in its place, then rate the straw man, a back door for inappropriately inserting editorial content into what's supposed to be a straightforward checking of checkable facts. As that Sanders example indicates, its analysts aren't always particularly scrupulous when it comes to sticking to that "checkable facts" part either. What's the point in conducting an investigation of a politician's assertion that he'd like to run against another politician? That Sanders wants to run against Trump definitely isn't checkable absent telepathic powers.

Over the years, I've written about some of the shortcomings of Politifact's work. What I'm going to outline here isn't, by any means, a comprehensive survey. Just some notes I and other critics have made on some of the things Politifact has done.

Back in 2011, I wrote a few things in various forums on behalf of Jon Stewart when he was straight-up mugged by Politifact. In an appearance on Fox News Sunday, Stewart had spoken about Fox viewers being misinformed. "Who are the most consistently misinformed media viewers? The most consistently misinformed? Fox, Fox viewers, consistently, every poll." The Politifact gang claimed this was "false." To "prove" it, they cited surveys that gauged how well-informed were consumers of various news outlets about various basic facts, surveys in which Fox viewers weren't the most uninformed. But Stewart didn't say "uninformed"; he said "misinformed." Every one of the surveys Politifact was citing to refute him asked respondents about general public affairs knowledge. Could they name the vice president? Could they name the president of Russia? And so on. Neither Stewart nor anyone else has ever claimed Fox lies to its viewers about things like that. Meanwhile, there had been survey after survey examining misinformation among news consumers and regular Fox viewers had proven the most misinformed in every one of them, no exceptions. In short, Politifact had gotten this flat-out, nowhere-to-run-or-hide completely wrong, but faced with a barrage of criticism, PF refused to back down or even deal with the issue.

And that's not unusual. Any press outlet is sometimes going to make mistakes--it comes with the territory. When PF's analysts get things wrong though, they routinely refuse to correct the error or only offer a half-assed correction or, perhaps more often than both, just dig in their heels and double down on the original error. It's the worst habit in the world for a "fact-checker."

(Stewart, btw, decided to "apologize" for his error--his reply, burying Politifact with its own work, proved why he's destined to be remembered as a comedy legend.)

In his Jan. 2012 State of the Union Address, Barack Obama had said, "In the last 22 months, businesses have created more than three million jobs. Last year, they created the most jobs since 2005." Politifact acknowledged that both statements were entirely factual but gave Obama only a "half true" rating because the checker said the president was insinuating his administration was responsible for this, something Obama had never said in the speech. The PF gang "refuted" this straw man they'd set up by referencing economists who noted that no presidential administration is entirely responsible for job growth in the economy. Another flurry of criticism followed and in a rare move, they did sort of back away--they changed their rating to "mostly true." For a statement they, themselves, had documented was entirely true in every particular.

They pulled this nonsense that same month with an Obama ad that said, "For the first time in 13 years, our dependence on foreign oil is below 50 percent." PF certified that this was entirely correct then rated the claim "half true" because, yet again, the analyst had read Obama's mind and asserted he'd taken credit for it but didn't deserve the full credit for it.

In February 2012, Politifact examined Sen. Marco Rubio's extravagant claim that "the majority of Americans are conservatives." This drew my attention at the time because it was a subject I'd repeatedly covered in the past. PF cited Gallup's most recent (at that time) self-identification poll which showed that 40% of Americans identified as "conservative" vs. 35% as "moderate" and 21% as "liberal." In short, by their own chosen source, 60% of Americans do not identify themselves as "conservative." Their rating of Rubio's claim? "Mostly true." I've written about the uselessness of such self-identification polls, perpetually subject to, among other things, how respondents interpret the well-traveled (and much-abused) words. When one looks at the absurdly extensive polling data available, one is hard-pressed to find any major issue on which Americans don't, by overwhelming margins, hold a liberal view. When their Rubio ruling was criticized, PF analysts changed their rating but not to "false," only to "half true." Apparently, PF's analysts have access to some special math by which 40% equals a "majority."

When, in 2012, Democratic Senate leader Harry Reid claimed he'd been informed by someone in the know that then-Republican presidential candidate Mitt Romney hadn't paid any taxes in at least a decade, it could be seen as a low blow, rumor-mongering, a dick move, political bullshit, or, alternately, some mischief Romney had earned via his refusal to publicly release his past tax returns (in the name of ethical transparency, presidential candidates traditionally release such returns). PF's Louis Jacobson, however, ranked Reid's comment  as a "pants on fire" lie. Jacobson asked some tax experts whether it was possible Romney had paid no taxes in a period of  a decade and they offered the opinion that this was "unlikely," which certainly doesn't rule out the possibility. In the real world, of course, there are only three grounds for that "pants on fire" rating: Jacobson had been given access to Romney's tax returns over an extended period, he'd had Reid under tight surveillance for an extended period or he had telepathic powers and could peer into Reid's mind. Jacobson doesn't even claim any of the three and absent them, there's absolutely no basis for PF's rating in the matter or, indeed, for any rating at all. Someone really could have said this to Reid. That "pants on fire" is strictly an editorial expression of disapproval of Reid's comments--again, not something an outlet that bills itself as a non-partisan fact-checker should be doing.

Perhaps the most infamous example of Politifact nonsense--because it provoked the biggest uproar--occurred in 2011 when Democrats had been noting that the Paul Ryan budget embraced by Republicans in the House of Representatives would end Medicare as we know it and PF decided that because the voucher program with which Ryan would replace Medicare would still be called "Medicare," this was false. On nine different occasions throughout 2011, PF rated iterations on it as either "false" or "Pants On Fire" lies, at a time when several proponents of the Ryan budget were publicly bragging it would end Medicare as we know it. PF analysts went on to compound this idiocy by choosing it as Politifact's "Lie of the Year" and when this provoked a wave of criticism, PF editor Bill Adair offered a follow-up in which he refused to address the facts and chose, instead, to take snotty ad hominem swipes at PF's critics, painting them as merely partisans living in an echo-chamber who didn't like independent fact-checkers.

That incident points to one of Politifact's editorial biases: it doesn't like "entitlement" programs. PF has, over the years, repeatedly embraced conservative mythology about such programs and has run interference for those attacking them.

The business about the Ryan budget was a sustained, long-running example of the latter. Another example of it: One PF article examined Indiana Rep. Todd Young's assertion that "nearly 70 percent of all federal spending will go towards Social Security, Medicare, and Medicaid" in the coming year (this was offered in 2013). PF documented that this was a false statement--those programs amount to only 45% of spending--and correctly rated it as "false." So far, so good. But then, the PF analyst spent over half the article trying to rationalize Young's comment, asserting that, if you add in all entitlement programs plus--utterly randomly--the full annual interest paid on the national debt, you get to 68% of spending--close to 70%. A more egregious example: In July, 2013, the Virginia Democratic party began running an ad against Republican gubernatorial candidate Ken Cuccinelli which asserted that in his new book, Cuccinelli "questions whether Medicare and Social Security should exist." Politifact's Sean Gorman examined this one and his "analysis" boggles the mind. Turning to the book in question, he quotes Cuccinelli:
"'Sometimes bad politicians set out to grow government in order to increase their own power and influence,' Cuccinelli wrote. 'This phenomenon doesn’t just happen in Washington; it happens at all levels of government. The amazing thing is that they often grow government without protest from citizens, and sometimes they even get buy-in from citizens -- at least from the ones getting the goodies.'

"Cuccinelli then cites specific entitlements.

"'One of their favorite ways to increase their power is by creating programs that dispense subsidized government benefits, such as Medicare, Social Security and outright welfare (Medicaid, food stamps, subsidized housing and the like). These programs make people dependent on government. And once people are dependent, they feel they can’t afford to have the programs taken away, no matter how inefficient, poorly run or costly to the rest of society.'"
That seems a pretty open-and-shut case. Virginia Democrats earn a "true" rating, right? Nope. Gorman instead turns to an entirely different part of the book 175 pages later wherein:
"Cuccinelli wrote that government spends little on investments that result in economic growth, noting that Social Security, Medicare and defense spending amounted to more than half of the federal budget in 2011.

"'There is no monetary return on these investments in any traditional business sense (that is, one invests money with a goal of getting a return in the form of interest, income or appreciation in value), although there are obviously other reasons America spends money on these programs,' Cuccinelli wrote. 'I’m not questioning here the existence of these programs nor the wisdom of how much money is spent on them. What I’m trying to illustrate is that most dollars that government spends do not create economic growth, but instead take money out of the hands of the people who do create economic growth.'"

Again, Cuccinelli wrote, 'I’m not questioning here the existence of these programs, nor how much money is spent on them.'"
Well, that settles that. Cuccinelli suggested that Medicare and Social Security were inefficient, poorly run, costly programs and flat-out said these were programs created by bad pols looking only to increase their own power but because he says, in an entirely different passage 175 pages later, that, in the context of that later passage, he isn't questioning their existence, he never questioned their existence at all and Politifact gives the Virginia Democrats' claim a rating of "mostly false."[1]

For years, Politifact has embraced false conservative claims that Social Security contributes to the federal budget deficit. In Sept. 2014, PF examined Oregon Sen. Jeff Merkley's assertion that SS "has never contributed one cent to the deficit. Not one cent." Though Merkley was correct, PF rated his statement as only "half true." This prompted a vigorous challenge from Jeff Hiltzik; writing in the Los Angeles Times, Hiltzik systematically dismantled PF's case, exposing a factual inaccuracy at the core of it and illustrating that PF was misrepresenting not only the SS program itself but even its own chosen source. PF analyst Dana Tims' response was to double down, misrepresent Hiltzik's criticism refuse to address the previous factual inaccuracies and add more errors to the pile. Hiltzik wasn't impressed. Politifact has clung to this same myth for years, despite having been corrected over and over again. In a way, the rating of "half true" given to Merkley could be seen as progress--in all of those previous iterations, PF analysts had ranked the same assertion as "mostly false."

In 2012, MSNBC host Rachel Maddow had started waging a recurring on-air battle against Politifact, routinely noting the kind of nonsense I've been outlining here. She has cited, for example, PF's attack on Obama for his remarks about jobs during his 2012 State of the Union Address, covered above, and PF's attack on former tennis champ Martina Navratilova's accurate assertion that "in 29 states in this country you can still get fired for not just being gay but if your employer thinks you are gay... We don't have equal rights." In the latter, Politifact's analyst Louis Jacobson documented the fact that "if you frame this statement in the context of blanket protections by states, she’s correct," then argued that, within the states that allow for such discrimination, there are some localities that don't, some employers that voluntarily pledge not to fire workers on such grounds and some government employees who enjoy protection against discrimination and while all of the above encompasses a microscopic fraction of the total workforce in those states--a fact PF's analyst left on the cutting-room floor--Jacobson decided they collectively conspire to reduce Navratilova's below "true," below "mostly true" and down to only "half true." Maddow roasted PF for it.

The Navratilova affair is classic Politifact. The org that makes so many appeals to "context" as rationale for a barrage of bad rulings takes a broadly true comment and attempts to make it hold the weight of the world in order to punch holes in it. What do the ludicrous extremes to which PF went to try to find some way to downgrade the truth of Navratilova's comment really say, other than that PF editorially disliked the fact that someone had pointed out homosexuals' lack of protection from discrimination in most states?

When it comes to dealing with an avowed critic like Maddow, all that hypertechnical concern for "context" goes right out the window. Back in late August, Politifact went after Maddow over something she'd said in mid-July. At the time, Wisconsin governor Scott Walker had just entered the presidential race and Maddow sized up his candidacy. At one point, she focused on Walker's unfortunate economic record. This is how Politifact presented this:
"'Under his [Walker's] leadership as governor, Wisconsin has had a really bad time of it when it comes to jobs and the economy,' Maddow said on her MSNBC program on July 13, 2015, the day Walker announced his run for the Republican presidential nomination.

"She added:

"'The Wisconsin middle class is shrinking at a faster rate than any other state in the country.'"
PF's Kevin Crowe and Tom Ketscher confirmed this statement was, in fact, true, using Maddow's own source, but they noted Maddow was using data that covered a 13-year period, including many years before Walker was governor. So they looked at middle-class shrinkage rates comparing one pre-Walker year to two Walker years and found that Wisconsin only ranked a still-pretty-awful 24th in the rate of decline in the middle-class during those Walker years. Any of the economists PF always makes a show of referencing could have (and would have, if any had been asked) told these analysts that unless there's some sort of dramatic, short-term change, a comparison of only three years of data is absolutely meaningless when it comes to measuring such a thing. The source Maddow cited was correct to examine it as a long-term trend. One can argue this makes it inappropriate for Maddow to cite this as emblematic of Walker's economic record and that criticism would certainly be valid. Maddow was given a rating of "false."

But PF, always that stickler for "context," isn't being honest here. While PF's presentation would lead an unwary reader to believe this factoid of questionable utility was the basis for Maddow's claim that, under Walker, Wisconsin has economically suffered and that PF had thus debunked that claim, that business about middle-class shrinkage was only part of a much longer monologue on the subject:
"Under his leadership as governor, Wisconsin has had a really bad time of it when it comes to jobs and the economy. Private sector job growth in his state is one of the worst in the nation. Wisconsin, the Wisconsin middle class is shrinking at a faster rate than any other state in the country. Over the course of his time in office, Wisconsin has steadily fallen in terms of job-creation numbers when ranked against other states. In 2011, they were 35th in the country, the next year they were 36th, the year after that, they were 38th. Last year, I should say, they held steady at 38th in the nation in terms of job growth. And under Scott Walker, Wisconsin's economy has done terrible vis-a-vis the rest of the country, vis-a-vis the region. It's done terrible compared with states of a similar size and a similar economic profile. If he does well in this giant Republican field, that economic record will probably be the thing that Democrats focus on when it comes to Scott Walker, his really bad economic record in Wisconsin."
PF's analysts didn't transcribe that and their silence on the rest of it--failing, in fact, to disclose there even was any "rest of it"--can probably be taken as a sign they didn't find anything else inaccurate in it and were reduced to ripping from context and nitpicking that single line, which, when left in place, is really of little consequence. Maddow's point was that Walker's economic record sucks and for a fact-checker, the story here is that she got this almost entirely right. What PF did has the appearance of a silly vendetta wherein the fact-checkers eschew their own stated principles in order to conduct a petty get-back at a critic.

Today's nonsense isn't the first time PF's analysts have issued bad rulings when it comes to Bernie Sanders. A much more egregious example happened back in August when right-wing columnist George Will, writing about the death of Robert Conquest, decided to throw in a slam at Sanders:
"Conquest lived to see a current U.S. presidential candidate, a senator, who had chosen, surely as an ideological gesture, to spend his honeymoon in the Soviet Union in 1988. Gulags still functioned, probably including some of the 'cold Auschwitzes' in Siberia, described in Conquest’s 'Kolyma.' The honeymooner did not mind that in 1988 political prisoners were--as may still be the case--being tortured in psychiatric 'hospitals.' Thanks to the unblinking honesty of people like Conquest, the Soviet Union now is such a receding memory that Bernie Sanders’s moral obtuseness--the obverse of Conquest’s character--is considered an amusing eccentricity."
The facts: Burlington and the Soviet city of Yaroslavl were to become "sister cities" under the old sister-city program established by Dwight Eisenhower to facilitate cultural exchanges between peoples around the world. Sanders, then the mayor of Burlington, was to lead a mayoral delegation that was to travel to Yaroslavl to make this official. Shortly before departing, he and his longtime girlfriend Jane O'Meara were married and she joined him as part of the Burlington delegation. PF's Will Cabaniss notes that this "'honeymoon' was dotted with meetings, interviews and diplomatic functions"--only a "honeymoon" in the sarcastic way the Sanders have referred to it over the years. He notes that the trip happened when the Gorbachev reforms were underway. He writes that while "Will made it sound as if Sanders was visiting to condone Soviet torture practices,.. the Burlington trip was more of a dialogue-building exchange program." Unmentioned by Cabaniss is that Sanders has never condoned torture or gulags. Did Sanders choose to honeymoon in the Soviet Union, as Will suggests, as an "ideological gesture"? Cabaniss deflates that one as well:
"When reached for comment, Sanders' campaign said that the dates for the trip had already been set, and the couple 'set their wedding date to coincide with that trip because they didn't want to take more time off.'"
So Will's comment was a lie in every particular, PF documented this and the rating for Will's claim--his libel of Sanders? "Mostly True." To get that, Cabaniss simply rated Will's comment about Sanders having, in fact, honeymooned in the Soviet Union and entirely ignored everything else, even after covering it.[2] As a final joke, Cabaniss wrote that "Will’s claim is accurate but is missing context..."

Politifact's work is often useful and the bulk of it may even hold up to scrutiny--I've certainly used it in the past--but it's riddled with this sort of nonsense and given the way so much of that nonsense breaks, one can't help but suspect a great deal of it is simply politically motivated--partisan politics constantly trumping any commitment to objective fact-checking. It's a remarkable--though unremarked upon--phenomenon that mainstream journalism has gone in such a direction that fact-checking is considered some sort of specialized art to be handled by specialized, segregated units and is no longer even assumed to be an integral part of the job of regular journalists. What it says isn't, in this writer's view, complimentary of the profession. One must, to be sure, exercise care in challenging an org that is endowed with this particular mission, especially in an atmosphere wherein the political right has waged such a long and relentless crusade against the very concept of an objective fact. The right-wing Rage Machine in the U.S. has instilled in its followers the belief that reality itself is subservient to their temporary political passions, that "liberal" means, in itself, "false" and down that road lies madness. At the same time, healthy, informed skepticism is, as always, a must. That such orgs make a show of dealing in facts makes it all the more important that they take great care to get it right when it comes to the controversies they're called to address. At this, Politifact has repeatedly failed. It's a fact-checker that needs a fact-checker. Take from it that which is of value and leave the rest.



[1] Two months after this absurd ruling, Politifact offered Cuccinelli another gift. His Democratic opponent Terry McAuliffe had created an ad noting that, if implemented, Cuccinelli's effort to roll back the state's no-fault divorce law would mean "a mom trying to get out of a bad marriage over her husband's objections could only get divorced if she could prove adultery or physical abuse or her spouse had abandoned her or was sentenced to jail." Cuccinelli's proposed bill had been part of a larger effort by right-wing "father's rights" groups to disadvantage women in such proceedings, an agenda Cuccinelli had embraced, but Politifact's Nancy Madsen ranked the ad "mostly false" on the absolutely absurd grounds that "in portraying the bill as an attack on women, McAuliffe ignores that the legislation would have made it equally more difficult for dads to get divorces." Both the Washington Post's Fact Checker and Slate offered detailed rebuttals of this absurd ruling.

[2] While only rating that one thin item, Cabaniss reproduced the entire Will libel against Sanders.

No comments:

Post a Comment