Torture narrative trumps facts at PolitiFact


Bill Adair, Duke University’s Knight professor of computational journalism, helped found PolitiFact. That likely helps explain his “Truth-O-Meter”-ish Twitter avatar.

Adair’s tweet about waterboarding helps show the way mainstream fact checkers help reinforce false partisan beliefs, in this case the popular narrative that torture does not work. And Adair accomplishes the feat despite the fact that his tweet is likely true: There is no solid evidence that waterboarding works. On the other hand, Adair’s statement does not settle the question of whether waterboarding works. And that is the key question.

Deceivers often do their deceiving by fudging the difference between two different but related statements. Adair’s tweet did not link to an article showing a lack of solid evidence that waterboarding works. His tweet led to a PolitiFact article calling “False” the statement that waterboarding works. The “ShareTheFact” image Adair tweeted makes that clear enough.

So what’s the problem, apart from Adair advertising the fact-check with either an inaccurate paraphrase or a tangential comment? This is the problem: The linked article provides no solid evidence that waterboarding doesn’t work, while at the same time sending the supposedly factual message that waterboarding does not work. That is deceptive. How does a fact checker justify drawing a conclusion without solid evidence?

We’ll parse the key parts of that article, along with another recent waterboarding-related PolitiFact fact check that similarly fudged the facts.

PolitiFact Florida vs Todd Wilcox

Todd Wilcox, a former CIA officer running in Florida as a Republican for a seat in the House of Representatives, said waterboarding works.

PolitiFact Florida ruled his claim “False” based on a string of bogus evidences.

Bogus evidence I

PolitiFact Florida offered evidence supposedly showing Wilcox was wrong, getting off on the wrong foot by citing “waterboarding” failures that failed to meet its own definition of waterboarding (“pouring water over a cloth covering their mouth and nose to simulate drowning”):

Waterboarding has long been considered a poor way to extract information, Reed College political science professor Darius Rejali said. Only a handful of case studies about waterboarding’s effects are available.

Those examples involved Nazi Germany’s Gestapo in Norway and France; the French in Vietnam and Algeria; and the United States in the Philippines at the turn of the last century. In each case, the benefits of waterboarding were suspect at best, Rejali said.


We have yet to find any description of Nazi “waterboarding,” let alone one that matches the definition PolitiFact Florida offers. The French “waterboarding” technique was not simulated drowning but literal drowning (suffocation). Note the description given by journalist  Henri Alleg, describing his experience in Algeria:

So, very quickly, the water ran all over my face. I couldn’t, of course, breathe. And after a few minutes, fighting against the impression of getting drowned, you can’t resist. And you feel as if you were drowning yourself. And this is a terrible impression of coming very near death. And so, when the paratroopers, the torturers, see that you’re drowning, they would stop, let you breathe, and try again.


Going for “a few minutes” without air is not a mere simulation of drowning. Recall that Khalid Sheikh Mohammed astonished interrogators by resisting almost two and a half minutes. Mohammed endured a very long treatment by U.S. standards. The treatment typically produced its results in seconds.

As for the “waterboarding” technique U.S. soldiers used in the Philippines, it filled the stomach with water, unlike modern waterboarding.

Testing the efficacy of techniques different from waterboarding serves poorly to test the efficacy of waterboarding.

We question the neutrality of the testimony from PolitiFact’s main expert source, Darius Rejali. Here is a paraphrase of Rejali from PolitiFact:

To start, interrogators aren’t in a position to know whether a prisoner is divulging anything factual, Rejali said. They think they have the training and can tell what’s the truth, but they can’t.


How does Rejali know that interrogators cannot tell whether a prisoner is telling the truth? Don’t interrogators ask questions to which they already know the answers as a basic technique for corroboration? We think that would provide a plausible way for interrogators to know whether they are getting the truth from a prisoner. We would like to see the evidence, if any, supporting Rejali’s claim.

So we have case studies of “waterboarding” that are different from U.S. waterboarding and no evidence supporting a claim that interrogators cannot tell when a prisoner is telling the truth. That is not solid evidence showing waterboarding does not work.

Bogus evidence II

PolitiFact uses a Senate report on waterboarding as a principal evidence waterboarding does not work (bold emphasis added):

The committee concluded that in each example, either key information was gathered before enhanced interrogation began, the information was already available prior to interrogation, or the interrogation had nothing to do with the information obtained.


Was the committee’s conclusion based on solid evidence? PolitiFact shows no signs of caring. The committee’s conclusion is served to the reader as evidence waterboarding does not work. But the report is weak evidence at best.

Bogus evidence III

PolitiFact cited a Newsweek story  supposedly showing waterboarding prevents a prisoner from providing reliable information.

Newsweek, citing Shane O’Mara, a professor of experimental brain research at Trinity College Dublin (bold emphasis added):

Like other enhanced measures, waterboarding cannot be tested in a laboratory for ethical reasons, but there is a sizable amount of relevant scientific literature on it. As O’Mara shows in his book, studies of the “diving reflex” (a set of physiological responses that occur when mammals, including humans, are submerged in water) have demonstrated that immersion in cold water moves brain activity away from areas supporting memory to those “principally concerned with survival,” such as the brainstem and amygdala, which regulate fear, pain and stress. By occluding the airways, waterboarding starves people of air, and there is a “huge literature” showing that lack of oxygen (hypoxia) harms cognition, O’Mara tells Newsweek. He highlights one recent study, which found that hypoxia “severely impairs” a person’s cognitive abilities. Furthermore, waterboarding causes carbon dioxide to accumulate in the body (hypercapnia), which induces fear and panic. In this situation, the ability to think and recall information will be “markedly reduced,” he says.


We do not know why O’Mara thinks the physiological effects of submersion in water (or immersion in cold water) carry relevance for the effects of waterboarding. Waterboarding does not involve submerging a prisoner in water. Nor does it involve immersing the prisoner in cold water. A room-temperature bottle of water would serve perfectly well for waterboarding as PolitiFact and Newsweek describe it.

We likewise do not know why O’Mara thinks hypoxia and hypercapnia carry relevance. As we noted above, a typical waterboarding lasts for mere seconds. Moreover, the technique is used to get an uncooperative prisoner to start talking. Prisoners who have started cooperating are allowed to breathe, relieving any problems with lack of oxygen or excess carbon dioxide.

The following video shows the late Christopher Hitchens experiencing waterboarding in 2008. Hitchens signaled “unbearable distress” in under 20 seconds. Hypercapnia in 20 seconds?

O’Mara’s claim that waterboarding reduces a prisoner’s ability to provide accurate information seems ill-founded. Any supporting experiment would need to closely match the conditions created by real waterboarding, not just O’Mara’s conception of waterboarding.

We do not know why the journalists writing these stories fail to make the same observations we are making, unless they carry some ideological bias that suppresses their tendency to question.

Bogus evidence IV

Supposedly if waterboarding worked, then it would be quickly adopted as an interrogation method by repressive regimes. No, really:

Rejali noted that along with the historical and scientific strikes against waterboarding, there’s a third element to consider. To put it bluntly, while much of the Senate’s report is still classified, there’s no proof that the CIA’s waterboarding program was effective enough to convince more brutal regimes to use it.

“If the Americans had found the golden fleece of torture techniques, no one has noticed and no one is copying,” Rejali said. The most effective methods of making someone talk are still torturing with electricity or beating them with sticks, he said. If waterboarding was considered truly effective, for example, “thousands of torturers in China would be using this. But they don’t.”


This argument is marvelously incoherent.

Rejali’s statement does not suggest that waterboarding is ineffective. Rather, PolitiFact’s expert argues that waterboarding is not effective enough to displace the “most effective” torture methods like electric shocks and beatings. So the expert says, sure, torture is effective but waterboarding isn’t as effective as more extreme techniques.

And that is all PolitiFact has. Four bogus arguments leading to a bogus conclusion. Todd Wilcox says waterboarding and other enhanced interrogation techniques work. PolitiFact finds his claim false because no proof exists that waterboarding works, and because PolitiFact absurdly buys the notion that there’s scientific proof showing it doesn’t work:

(T)here’s scientific proof that a technique like waterboarding would affect brain function enough to make any prisoner’s statements unreliable. They may say anything to make the waterboarding stop, and could actually be physically unable to provide any cogent intelligence.


There’s no scientific proof that waterboarding as PolitiFact describes it—and we think it used an apt description—would lead to the consequences PolitiFact describes above. PolitiFact Florida’s conclusion misrepresents and wildly exaggerates the evidence discussed in its fact check. PolitiFact Florida did not uncover enough evidence to answer whether waterboarding works, yet ruled waterboarding does not work.


PolitiFact and Hillary Clinton


Whereas PolitiFact Florida took an adversary position against Republican Todd Wilcox, PolitiFact National did what it could to make a false claim from Hillary Clinton seem true.

Clinton said we know torture doesn’t work “based on lots of empirical evidence.” PolitiFact ruled that claim “True” on March 30, 2016. I wrote about it for the PolitiFact Bias blog, noting “We do not possess enough empirical data to know torture does not work. Giving Clinton a “True” rating makes a total mockery of journalistic objectivity. ”

What evidence did PolitiFact find to back Clinton’s claim?

PolitiFact started off by giving away the store, conceding that torture mostly evades scientific study since it requires torturing research subjects. After that, PolitiFact started with the same type of bogus evidence PolitiFact Florida used in its fact check.

Bogus evidence A

The experts with whom we spoke said that, for a host of reasons, torture has long been recognized as an inefficient way to get information.


PolitiFact uses a survey of experts to help support the point that we know torture doesn’t work. But PolitiFact’s reporting does not show the experts saying torture does not work. The experts, PolitiFact says, say torture is inefficient. Is that supposed to be the same thing?

Bogus evidence B

If tortured subjects say many untrue things, does that allow us to know torture doesn’t work? What percentage of true statements would foreclose the hypothesis that torture does not work? Five percent? One percent? Nonetheless, PolitiFact takes that path:

“It is enormously well documented (including by the Senate study of the CIA program) that torture generated enormous amounts of false statements,” said Lisa Hajjar, a professor at the University of California, Santa Barbara, who studies torture.


We do not see how enormous numbers of false statements would counteract the conclusion that torture works if torture can produce at least one true statement. We could enlist Hajjar’s expert opinion in support of the claim that torture is inefficient, of course. But we are not convinced that knowing torture is inefficient is the same thing as knowing torture does not work.

Bogus evidence C

We covered this evidence under “Bogus evidence I” in our previous section. Once again, we have Darius Rejali saying interrogators can’t tell when a tortured prisoner is telling the truth:

Interrogators aren’t particularly good at telling when someone under torture is lying or telling the truth.

“Interrogators in these situations think they’re perfect lie detectors but research tells us quite differently,” said Darius Rejali, a torture expert at Reed College in Oregon.


We’re not sure what research Rejali can use to judge CIA interrogators. PolitiFact does not reference the research, so far as we can tell.

We reached out to Rejali via email to ask how he reconciles this claim with his claim, reported by PolitiFact, that China would use waterboarding instead of electric shock and beatings if it worked better. Are interrogators who rely on beatings better at judging truth than interrogators who use waterboarding? Or what?

Rejali’s representative at Reed College received our message. We will update this item if we receive a reply from Rejali.

Update: On June 2, 2016 we received a response from Rejali (available through the hotlink just above). Rejali said he noticed that PolitiFact’s paraphrase of his statements left a paradoxical impression and he asked for a clarified version. PolitiFact updated its fact check with new wording and a clarification notice dated June 1, 2016.

The new wording of Rejali’s comments does nothing to improve the point as evidence waterboarding doesn’t work. The rephrasing emphasizes the effectiveness of electric shock and beatings in getting prisoners to talk, but not necessarily to provide true information.  It sidesteps the issue of whether the techniques might help collect valuable intelligence information.

Bogus evidence D

PolitiFact says public opinion can work against proponents of techniques like waterboarding. We’re not entirely sure we should count this as PolitiFact’s justification for saying there is empirical evidence torture does not work. But we don’t want to leave it out, just in case PolitiFact did intend it as an evidence.

Another issue is that news reports of torture turn public opinion against authorities who condone or permit torture.


If torture works, it works regardless of public approval.

Bogus evidence E

Bogus evidence E overlaps Bogus evidence II from the Wilcox fact check.

The Clinton fact check offers a rambling rundown of a Senate committee report on enhanced interrogation. The Clinton fact check offers a different summary of the committee’s work than we see in the fact check of Wilcox:

The committee’s analysis of 20 prominent cases that reportedly “saved lives” — cited by the CIA and the Bush administration — found either that the key information was gathered before torture began, the information was already available to the intelligence community before the enhanced techniques were initiated, or there was no relationship between those success stories and the information provided by the detainees.


In both of its condensed versions of the Senate committee’s majority findings, PolitiFact mentions that information obtained after enhanced interrogation was already available to the intelligence community. This should not obscure two facts: The interrogation produced accurate information, and interrogators were able to identify that accurate information at least partly through corroboration. Interrogation after enhanced techniques produced new prisoner testimony corroborating existing information. We mentioned this earlier as a basic method for finding whether information is true. The Senate committee spins corroboration as a weakness. But corroboration helps interrogators focus more appropriately on the most reliable information coming from prisoners.

PolitiFact serves up the Senate committee’s spin without questioning its basis in fact.

PolitiFact gives us Rejali’s two cents in this section, also:

The problem with most of the accounts of people who say that torture works, said Rejali, is that only the cases where it seemed to succeed are remembered and the ones where it didn’t are forgotten.


To us, receiving Rejali’s opinion through PolitiFact’s paraphrase, Rejali seems to be saying that if a Major League designated hitter is batting .051, then hitting a baseball with a baseball bat doesn’t work. We can’t ignore, he says by our analogy, the batter’s high failure rate. We think this argument fits okay when arguing that torture works inefficiently. But we think the argument does not work in support of the claim that torture does not work at all.

PolitiFact ultimately used the fallacy of appeal to ignorance to justify giving the Republican, Wilcox, a “False” rating.

Gary N. Curtis, the creator of the website, describes the fallacy:

An appeal to ignorance is an argument for or against a proposition on the basis of a lack of evidence against or for it. If there is positive evidence for the conclusion, then of course we have other reasons for accepting it, but a lack of evidence by itself is no evidence.


PolitiFact cut the Democrat, Clinton, a break by ignoring the substance of her claim. Instead of looking for the “empirical evidence” that would allow us to know torture does not work, PolitiFact effectively removed the burden of proof from Clinton and gave her a “True” rating. Making a weak case against waterboarding with a series of bogus evidences was enough for PolitiFact to find it “True” that empirical evidence shows waterboarding does not work.

That brings us back to Bill Adair. We think Adair is right that no solid (scientific) evidence exists proving that waterboarding works. But using that lack of evidence to justify spreading the story that torture does not work repeats the appeal to ignorance fallacy.

In truth, PolitiFact’s fact check provides some evidence, well short of proof, that waterboarding works. And PolitiFact utterly failed to verify Clinton’s claim that we have evidence allowing us to know torture does not work.

We don’t have the evidence to know whether waterboarding works. PolitiFact had no business issuing “True” and “False” ratings of statements it cannot verify as true or false. At least not while pretending to serve as a neutral fact checker.

Bias serves as the best explanation for PolitiFact’s treating the two claims according to different standards and justifying its ratings fallaciously.

Bill Adair helps spread that misinformation with Duke’s newfangled “ShareTheFact” widget.

Leave a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.