Monday, January 1, 2018

PolitiFact's worst 17 fact checks of 2017

PolitiFact had a terrific year churning out horrible fact-checking content. In 2016 we published a list of PolitiFact's 11 worst fact checks. This year we're expanding that number to 17. Some of the selections show off PolitiFact's tendency to check statements of opinion. Some of them focus on other methodological blunders. And some make our list based on their potential impact on public debate.

Let's get to it.


17) PolitiFact Georgia says lawyers win a case if it's settled out of court!


Stacey Evans, a Democrat running for the governorship of Georgia, claimed she "helped win one of the biggest private lawsuits against Medicare fraud in history." PolitiFact found the claim unreservedly "True." The problem? The parties settled out of court.

To be sure, the amount recovered in the settlement was very large: over $300 million after subtracting considerable attorneys' fees. But winning a settlement simply counts as different from winning a lawsuit. If settling counts as a win, then attorneys on both sides might claim victory. Victory in court does not carry that type of ambiguity. PolitiFact Georgia simply let the imprecision slide.


16) PolitiFact cannot say whether Joy Reid was right about white nationalists in the White House


When liberal pundit Joy Reid charged that President Trump's White House staff included white nationalists, PolitiFact found it convenient to toss its "burden of proof" principle out the window and print an article suspending judgment on Reid's claim. We suppose that guilt by association was enough for PolitiFact to justify its move into "We Report, You Decide" territory.


15) Roy Moore & family took over $1 million from a charity while receiving pay as a judge?


In finding it "Mostly True" that Roy Moore took over $1 million from a charity while working for the state of Alabama as a judge, PolitiFact provided a landmark example of losing focus on the important point.

It is perfectly okay to "take" over $1 million from a charity, particularly if you are doing work for that charity, the pay rate falls in line with the standard for that type of work and the total represents years of such work. On the other hand, taking that pay while working as a judge looks suspicious.

PolitiFact found Moore did not take the pay while working as judge, but PolitiFact pasted a "Mostly True" graphic next to the claim that he did.

If PolitiFact noticed the criticism it received on this point, it did nothing to fix it:




14) The profound "Mostly True" meaning of a meaningless gender wage gap statistic

Mainstream media fact checkers tend to turn a blind eye on the use of the raw gender wage gap to imply a large pay gap stemming from discrimination, and PolitiFact is no exception. Note the "Truth-O-Meter" graphic from PolitiFact Florida:


Democratic Women's Club of Florida vice-president Patricia Farley said the raw wage gap in Florida, 79 to 80 cents on the dollar, happens to a woman "simply because she isn't a man." That language strongly implies gender discrimination as the cause of the entire raw wage gap. Mainstream fact checkers know full well (or should) that is not the case but repeatedly fail to call out the deception.


13) PolitiFact California uses stats for all immigrants to argue undocumented immigrants commit crimes at a lower rate than native-born Americans

In order to exhibit irresponsibility, a fact checker may use stats that measure one thing to support a conclusion about a separate thing. Undocumented immigrants are not all immigrants, and it counts as completely unscientific approach to assume the same crime rates for legal immigrants as for undocumented ones.

The study PolitiFact relied on to support the claim as "Mostly True" appeared vulnerable to other criticisms as well. Prison populations do not necessarily mirror  demographics for the general population, for example.

12) New York achieved "pay equity"?

PolitiFact New York makes our list for proclaiming it "Mostly True" that the state of New York has achieved gender pay equity. Lt. Gov. Kathy Hochul made the claim while also stating that in New York women make 90 cents for every dollar a man earns.

Everyone feeling the equity?

PolitiFact New York covered for its absurd ruling by claiming that Hochul's underlying point was true. The fact checkers claimed Hochul was making the point that New York has a significantly smaller gap than the rest of the country.

PolitiFact did Hochul a huge favor by not quoting her further: "Now it's got to get to a hundred." Why, if the reasons for the pay gap were as PolitiFact New York described them (different jobs, taking off work to care for a family, etc.) do women need to earn the same as men? Pay equity sounds reasonable if the gap stems from gender discrimination. Otherwise, not so much.

11) Bernie Sanders said a GOP health care law would throw 23 million off their insurance? So what?

When Sen. Bernie Sanders said a proposed GOP health care bill would throw 23 million off their health insurance leading to the death of thousands, PolitiFact entirely ignored the claim about the 23 million to focus on the thousands who would die as a result of losing their insurance ("Mostly True," said PolitiFact).

PolitiFact appeared to simply take Sanders' claim at face value. On the other hand PolitiFact's June 22 explainer article stated as fact "The CBO estimated 23 million people would lose insurance under the House health care bill." How's that for consistency?

10) PolitiFact says white men commit the majority of mass killings in the United States!

Though it is true that white men commit the majority of massing killings in the United States, we count PolitiFact's Oct. 6, 2017 fact check as one of our worst for what it ignores and fails to spell out for its readers.

PolitiFact was checking a claim ready-made for bashing white men but it failed to emphasize how the data undercut that purpose. Men commit the vast majority of mass shootings, and white men carry responsibility for roughly their share in terms of race. Using the data as an attack on white men was a non-starter, but PolitiFact whiffed on emphasizing the point. The data show men are far more likely than women to commit mass shootings. The notion that whites commit more than their share of the mass shootings gets no support from the data.

PolitiFact often rates true statements "Half True" or worse if, for example, it judges the true data point as "meaningless." This meaningless data point skated with a "Mostly True" rating.

9) PolitiFact says collusion, by itself, is a crime!

When Fox News legal expert Gregg Jarrett made the argument that collusion by itself does not count as criminal (except as an antitrust violation), PolitiFact rated his claim "False." PolitiFact supported its rating by citing experts who pointed out that if the collusion was part of some other crime then it was criminal. Derp. It's always great when experts match PolitiFact's poor attention to detail.

It makes no sense at all to argue that because collusion plus x is a crime therefore collusion by itself counts as a crime. Experts should know better, though PolitiFact may have thrown them off with leading questions or a misleading presentation of Jarrett's argument.


8) PolitiFact says conjecture equals assertion of fact for Betsy DeVos

PolitiFact Bias co-editor Jeff D. spotted this item. During the confirmation hearing for Education Secretary Betsy DeVos, DeVos was asked whether guns have any place in or around schools. Reference an exchange earlier in the hearing, DeVos said she imagined that the school in Wyoming that has a fence for keeping grizzly bears out of the school has a gun on hand to protect students from bears.

PolitiFact took DeVos to say that the school has a gun on hand to protect against grizzly bears. As the school did not have a gun on hand for that purpose, PolitiFact ruled DeVos' conjecture "False."

Isn't it obviously improper to treat conjecture as a statement of fact?

7) Was it 17 agencies or merely three or four?

PolitiFact rated "True" Democratic presidential candidate Hillary Rodham Clinton's October 2016 claim that 17 agencies "all concluded that these espionage attacks, these cyberattacks, come from the highest levels of the Kremlin, and they are designed to influence our election." It turned out that at least 12 of the agencies didn't even look at the question, but PolitiFact reasoned that since the Department of Homeland Security and the Office of the Director of National Intelligence represented those 17 agencies and had offered the judgment therefore all 17 agencies made the finding.

No, we're not kidding. That was how PolitiFact reasoned. And PolitiFact absurdly defended that bogus reasoning in a separate article in July 2017.

Quick summary: It's okay to present the finding, as DHS and the director did, as a finding of the U.S. intelligence agencies. Specifying the number of agencies under that umbrella regardless of whether they looked at the issue counts as deceptive. No credible fact checker could overlook that.


6) The CBO thinks repealing the individual mandate cuts Medicaid enrollment? Who knew?

PolitiFact's spirited defense of the Affordable Care Act included botched reporting on GOP proposals to overhaul the ACA. Ramesh Ponnuru wrote a piece criticizing PolitiFact for failing to report the CBO's judgment that repealing the individual mandate would result in millions fewer signing up for Medicaid. The CBO said the mere threat of the mandate tax prompted people to sign up for Medicaid who would not otherwise do so. PolitiFact omitted that from its reporting and even suggested as a blanket statement that lower projections for Medicaid enrollment counted as involuntary loss of insurance.

We've needled PolitiFact repeatedly on this point for months, and PolitiFact has yet to correct its past omissions and/or false reporting.


5) Courts use the "clear and convincing evidence" standard in criminal trials?

Lawyer and persistent PolitiFact critic Hans Bader caught PolitiFact Pennsylvania claiming that criminal courts use the "clear and convincing evidence" standard. The fact checkers were looking at claims about proposed changes to the Obama administration's guidance for Title IX issues, which coerced colleges into using the "preponderance of evidence" standard for sexual assault cases handled on-campus.

We estimate that most Americans know that prosecutors must demonstrate guilt beyond a reasonable doubt for a criminal case. PolitiFact partially corrected its error, but replaced it with an equally severe error by falsely claiming that the Foundation for Individual Rights in Education sought to let colleges use the "beyond a reasonable doubt" standard for sexual assault cases. In fact, FIRE advocated letting colleges use the "clear and convincing evidence" standard. PolitiFact Pennsylvania's error remains to this day.

The fact checkers ended up sending a very misleading message on precisely the issue they tried to address.


4) PolitiFact's 2017 "Lie of the Year" was a straw man fallacy

A straw man fallacy occurs when one misrepresents the argument of another by changing it to a weaker and easier-to-attack form. PolitiFact appeared to do that with its 2017 "Lie of the Year" selection, interpreting President Trump's statement that "this Russia thing with Trump and Russia is a made-up story" to mean the Russia election interference story minus Trump was made up.

It seems to us obvious that referring to "this Russia thing with Trump and Russia" has to include Trump to avoid committing the straw man fallacy. We challenged PolitiFact to produce clear evidence Trump made the claim it awarded its "Lie of the Year" award (no response so far).

How does the outfit that created the "Lie of the Year" fact-checking award end up botching it so supremely?


3) Democrats say the GOP health care reform bill will effectively end Medicaid? So what?

Remember back when PolitiFact awarded its 2011 "Lie of the Year" to the Democrats' claim that the GOP would end Medicare? Maybe if Democrats like Senator Kamala Harris (D) claimed the GOP would effectively end Medicaid, PolitiFact would fact check the claim?

The claim received an evaluation from PolitiFact California but no "Truth-O-Meter" rating. However, the article did make a key concession: "Harris’ statement could be construed as saying Medicaid, as it now exists, would essentially end."

We're still wondering how one could construe "effectively will end Medicaid" other than Medicaid, as it now exists, would essentially end.


2) Odds are PolitiFact California does not understand probability

PolitiFact California fact checked whether the yearly risk of an individual dying in America from a terrorist attack conducted by a refugee was 3.6 billion to 1. PolitiFact found it "True." And that would be fine, except PolitiFact California has taken its fact check to mean that the yearly risk of a fatal terrorist attack conducted by a refugee is 3.6 billion to 1. Making that mistake offers a convincing argument that PolitiFact California has no clue about probability calculations.

Using the data PolitiFact California used, the probability of a fatal attack occurring come out to 13.3 to 1. That's a huge difference, to say the least.

PolitiFact California presumed to explain to others what it clearly failed to comprehend.


1) PolitiFact's partisan evolution on budget cuts from a projected baseline

Zebra Fact Check documented how all the so-called "elite three" fact checkers, FactCheck.org, PolitiFact and the Washington Post Fact Checker experienced a type of epiphany on budget cuts from a projected baseline. Under Obama the fact-checkers found that a cut from a future baseline wasn't really a cut if budgeted funds nonetheless increased. Under Trump, the fact checkers spontaneously discovered that it was fair to call a cut to a future projected baseline a cut, even if the annual budget increased every year.

PolitiFact's version of the shifting standard makes No. 1 on our list because of its brazen misrepresentation of its own history. PolitiFact falsely claimed that it had generally rated GOP claims the ACA cut Medicare as "Half True." Nearly three quarters of PolitiFact's ratings on the topic were "Mostly False" or worse.

PolitiFact supported a biased narrative on budget cuts and hid its bias with false reporting.

No comments:

Post a Comment

Thanks to commenters who refuse to honor various requests from the blog administrators, all comments are now moderated. Pseudonymous commenters who do not choose distinctive pseudonyms will not be published, period. No "Anonymous." No "Unknown." Etc.