A source is a source, of course, of course

Senator Harry Reid’s office recently published a bit of fiction that claimed “in the history of the United States, 168 presidential nominees have been filibustered; 82 blocked under President Obama, 86 blocked under all other presidents.” PolitiFact produced an opinion piece that failed to find much wrong with this statement. As is all too typical, the article was labeled a “fact check” while it bypassed actual fact-checking opportunities.

Reid’s office responded to PolitiFact’s request for supporting documentation with two pertinent documents from the Congressional Research Service. PolitiFact honed in on the key issue in one of them:

The most recent of the two documents, a CRS memo, said, “In brief, out of the 168 cloture motions ever filed (or reconsidered) on nominations, 82 (49 percent) were cloture motions on nominations made since 2009.”

To the uninitiated, that might seem to support Reid’s claim. PolitiFact clearly believes it does. But the astute reader will notice that the claim is that nominations were “filibustered” and “blocked.” The CRS says that cloture motions were “filed (or reconsidered).” If cloture motions were always filed in response to filibusters, Reid’s office might be correct. But, as the Senate majority leader surely knows, that’s not true.

Another document from the Congressional Research Service explains the nuances:

Although cloture affords the Senate a means for overcoming a filibuster, it is erroneous to assume that cases in which cloture is sought are always the same as those in which a filibuster occurs. Filibusters may occur without cloture being sought, and cloture may be sought when no filibuster is taking place. The reason is that cloture is sought by supporters of a matter, whereas filibusters are conducted by its opponents.

It is possible, as a result, that opponents of a matter may use debate and other procedural actions to delay a vote without supporters deciding to move for cloture. This situation appears not to be common today, but does seem to have occurred in relation to nominations in earlier times. Supporters may refrain from seeking cloture either because they think they lack the votes to obtain it, because they believe they can overcome any delaying actions and reach a vote without resorting to cloture, or because they hope to resolve the matter in dispute by some negotiated accommodation.

On the other hand, leaders of the majority party, or other supporters of a matter, may move for cloture even when opponents deny that they are conducting a filibuster, or at a point when no extended debate or delaying actions have actually occurred. They may do so in response to a threat or perceived threat of a filibuster, or simply in an effort to speed action. Under contemporary conditions, in particular, it has often appeared that Senate leaders attempt to avoid bringing to the floor matters, including legislation as well as nominations, on which they foresee a likelihood that filibusters will occur. These agenda choices may be motivated in part by a desire to avoid expending scarce floor time on matters that cannot be brought to a successful conclusion.

Cloture motions are clearly not a proxy for filibusters, but PolitiFact accepted the erroneous conflation of the two without note. One might attribute their gaffe to the Dunning-Kruger Effect (i.e. they are not as skilled in understanding Senate procedures as they believe themselves to be). The issue with that theory is that the preceding quote comes from Cloture Attempts on Nominations: Data and Historical Development, one of the documents supplied by Reid’s office to “support” the claim—and even more embarrassingly, one of the documents PolitiFact listed as a source. The text appears in a section titled, in big bold type, Cloture Motions Do Not Correspond With Filibusters.

Harry Reid is well aware that there haven’t been 82 filibusters of President Obama’s nominations, but made the claim anyway. When asked for supporting documentation, Reid’s office actually provided documents that refuted the claim: “it would be a misuse of the following data, which identify nominations on which cloture was sought, to treat them as identifying nominations subjected to filibuster.” That’s the kind of disingenuity that a fact-checking operation should expose. But, if you want to be a fact checker, you have to be willing to actually check the facts.

Edit November 25, 2013 at 6:53 am: Added missing quote from CRS to concluding paragraph.

By their own standards

Politifact recently passed judgment on a statement by Senator Bernie Sanders (I-VT): “One out of four corporations doesn’t pay a nickel in taxes.” This statement refers to the fact that approximately three out of four corporations pay federal income taxes in any given year. Politifact noted that the reason that many corporations pay no taxes is due to losses carried forward from previous years and therefore assigned a “Half-True” rating to the statement.

The opinion offered by Politifact seemed unremarkable, and it might be, if it weren’t for their assessment of this statement by Congresswoman Michele Bachmann: “The top 1 percent of income-earners pay about 40 percent of all taxes into the federal government.” Bachmann was referring to the fact that approximately two out of every five dollars in federal income taxes is paid by the top one-percent of earners. Despite this being pretty accurate, Politifact assigned a “False” rating. They justified their rating thusly: “Bachmann would have been right if she’d said, ‘the top 1 percent of income earners pay about 40 percent of all income taxes into the federal government.'”

It seems that, according to the Politifact rating system, failing to use the appropriate modifier for taxes damns ones statement. It is taken from the promised land of “True” to the perdition of “False”. Politifact is certainly entitled to formulate their opinions using standards of their choosing. But, clearly, the same standard was not applied to the statement by Bernie Sanders. Despite its also lacking the income modifier, the “Half-True” rating was assigned.

Politifact assured its readers that Sanders was referring to income tax–an assurance they would not offer for Bachmann despite the appearance of the word income in her statement. Politifact justified this assurance by misinforming its readers that Sanders made his statement as he and Senator Lindsay Graham (R-SC) “were talking about federal income taxes.” Politifact would have been right if they had said “medical device taxes.”  The discussion transcript plainly shows that the tax issue under discussion was the repeal of the medical device tax, not income tax. In fact, use of the income came much later when Senator Graham discussed means testing for Social Security.

If a “False” rating for improper modification of taxes is proper in Bachmann’s case, it should also not only apply to Sanders, but to Politifact as well for their review of the Sanders statement. If a “False” rating applies to the adjectival slips of others, it must also be considered a proper assessment for Politifact’s self-same bungling.

Jeepers

PolitiFact and Glenn Kessler have a joined the chorus of voices taking the Romney campaign to task for an that criticizes Chrysler for shifting Jeep production to China. PolitiFact and Kessler disparage the ad. Both rely, in part, on the following statement from Chrysler’s headquarters (emphasis Kessler’s):

Jeep has no intention of shifting production of its Jeep models out of  North America to China. It’s simply reviewing the opportunities to return Jeep output to China for the world’s largest auto market. U.S. Jeep assembly lines will continue to stay in operation. A careful and unbiased reading of the Bloomberg take would have saved unnecessary fantasies and extravagant comments.

PolitiFact concludes that the ad is erroneous because it “leaves the clear impression that Jeeps built in China come at the expense of American workers.” Kessler reaches a similar conclusion saying that “the unspoken message is that American jobs are being sent to China, even though the ad carefully tiptoes around that claim.”

It seems pretty simple doesn’t it? The Romney ad claims that American workers will lose their jobs when Jeep production moves to China. Chrysler implies otherwise. Therefore, one can conclude the ad is false — at least if one is a non-fact-checking “fact checker.” An actual fact checkers might see the disagreement between Romney and Chrysler views to — I don’t know — perhaps, check the facts?  (It’s should be no surprise that PolitiFact leapt to its conclusion without doing any actual fact checking. But Kessler disappoints here as he usually puts forth an actual effort to check facts.)

It is clear Chrysler is moving the Jeep production to China is to produce vehicles for the Chinese market. The statement from Chrysler that “U.S. Jeep assembly lines will continue to stay in operation” is being used to refute the advertisement’s assertion that “Mitt Romney will fight for every American job.” Both Kessler and PolitiFact reject the implication that U.S. jobs are in danger. This conclusion relies upon the assumption  that no U.S. jobs will be lost when Jeep production for China moves to China. That is a very bad assumption.

The Jeeps currently sold in China are produced in the U.S. In fact, lawmakers from Michigan have pressured the Obama administration to crack down on the Chinese tariffs on these vehicles. The requested crackdown has had little impact. But, Chrysler can, and will, avoid the tariffs by producing Jeeps in China using Chinese workers.

For those slow on the uptake: Jeep vehicles that will be produced in China by Chinese workers are currently produced in the U.S by U.S. workers

It is Chrysler’s insinuation that a move will not affect U.S. workers that is disingenuous, not Romney’s indication to the contrary. When Jeep production for the Chinese market moves to China, it will reduce the number of jobs in the United States and result in the need for fewer U.S. workers. Those are the facts — for anyone who cares to check.

Tax evasion

PolitiFact recently fact checked the following statement: “Win Olympic Gold, Pay the IRS: U.S. Olympic medal winners will owe up to $9,000 to the IRS.” The statement is the subtitle of an article by Americans for Tax Reform (ATR). PolitiFact assigns a “Mostly False” rating saying: “it’s not likely that anyone would pay that much per medal in taxes — even if the winner was fortunate enough to have annual income well over $380,000 and refused to deduct any business expenses on their winnings. Any accountant worth their salt should be able to get the rate of tax on medal winnings much below $9,000, and maybe even to zero.” There are errors that resulted in the aforementioned rating, but the errors are PolitiFact’s, not ATR’s.

The $9,000 figure comes from a straight-forward calculation. ATR estimated income from winning a gold medal by totaling the approximate commodity value of a medal of $675 together with the $25,000 U.S. Olympic Committee bonus for gold medal winners. Next they applied the top marginal tax rate of 35% to the $25,675 and determined that a gold medal winner in that tax bracket would owe an extra $8986 in taxes. These are the simple mathematics that result the claim that “winners will owe up to $9,000 to the IRS.”

PolitiFact’s ruling makes it clear that they have difficulties with even simple mathematics. When they state that “it’s not likely that anyone would pay that much per medal in taxes,” they conflate an average (“per medal”) with a maximum (“up to”). Their next sentence discusses “get[ting] the rate of tax on medal winnings much below $9,000.” Of course, $9,000 is a tax amount, not a tax rate. If there was a good undestanding of tax issues at PolitiFact, these errors likely would not have appeared. Because these misstatements formed the core of PolitiFact’s ruling, it can’t be regarded with any confidence.

ATR’s statement and calculations are correct. Each athlete in the 35% tax bracket that reports $25,675 in income from winning a gold medal will owe an extra $8986 in taxes. Despite what PolitiFact seems to imply, there are no tax deductions given to those that win gold medals that aren’t afforded to those who finish last. A medal winner’s marginal tax rate will usually determine the amount of the extra taxes paid. Deductions will come into play only for those with minimal athletic income.

A poor choice of words

Hazel Sheffield ofthe Columbia Journalism Review recently posted an item, “The heatwave debate“, about journalism and climate science. One paragraph discussed the use of probabilities in science reporting. Here is that paragraph sans its final clause:

 Using probability to calculate the likelihood that a weather event was caused by global warming is a new science, known as “probablistic event attribution”. But it’s already changing the way journalists talk about climate change, by lending them the scientific habit of never being definitive. That can have a damaging effect on the semantics journalists use to describe freak weather events, as David Roberts notes in a post on Grist. For example, we routinely say, “Smoking causes lung cancer,” even though, in some cases, smoking might not be the cause. We should be equally careful of our word choices when reporting on climate change, or…

The next clause is the final one in the paragraph. It should bring logical and thematic closure. It could warn against portraying probabilities as absolutes. It could caution against using improper terminology. It could… Well, it doesn’t matter what it could do. Here is what it did:

We should be equally careful of our word choices when reporting on climate change, or skeptics can seize on the ambiguities.

Careful word choice is often extolled as the path to precision and clarity. Here it is being offered as the way to deny rhetorical ammunition for a certain point a view. A point of view that the author, presumably, does not share. There is a dissonance between this call for partiality and the use of the word reporting.

So, the next time someone tells you that there is no media bias, or that there is media bias but that it’s unintentional, or that there is intentional media bias but that it’s isolated, you can tell them that one time an editor from the Columbia Journalism Review called on reporters to engage in premeditated bias against a disesteemed viewpoint. They’ll never believe you.

Without Our Complements

PolitiFact reviewed a statement from an Obama campaign ad that said “Mitt Romney’s companies were pioneers in outsourcing U.S. jobs to low-wage countries.” In their ruling PolitiFact stated: “We find little evidence that the particular firms were ‘pioneers in outsourcing.'” Despite the lack of evidence, they blundered into assigning a “Half True” rating to the statement.

The ruling was the result of curious decision by PolitiFact:

We decided to fact-check two parts of the Obama campaign’s statement. Were these really Romney‘s companies? And were these companies actually ‘pioneers in outsourcing’?

The Obama campaign’s statement is simply a copular sentence. There’s the subject: “Mitt Romney’s companies”; the linking verb: “were”; and the subject complement: “pioneers in outsourcing U.S. jobs to low-wage countries.” PolitFact decided to fact check the subject of such a sentence while disregarding the subject complement. That is, they decided to fact check the phrase “Mitt Romney’s companies” and then found that to be accurate–hence, the “Half True” rating.

I suppose one could assign the same rating to the statement, “PolitiFact’s articles are full of nothing but falsehoods,” because the articles are, after all, PolitiFact’s. But one could only do this by committing the same error that PolitiFact did. You can’t verify the subject of the sentence while ignoring the subject complement. Defining the relationship between the two is the whole purpose of a copular sentence.

The Obama campaign’s statement should not have been difficult to analyze. PolitiFact found the evidence to refute it. The analysis went off the rails with the botched decision to validate the phrase “Mitt Romney’s companies.” That led to the erroneous “Half True” ruling.

A good understanding of English grammar and usage is critical to quality fact checking. It’s dispiriting to see PolitiFact blunder with such a basic type of sentence.

Fail

PolitiFact analyzed the following statement by Robert Gibbs: “Nobody knows why [Mitt Romney] has a corporation in Bermuda, why he failed to disclose that on seven different financial disclosures.” After introducing the issue, PolitiFact commended two articles: “Reporting by the Associated Press and Vanity Fair teed this up. Both news organizations explored one of Romney’s overseas accounts, Sankaty High Yield Asset Investors, Ltd.”

Both articles document the absence of Sankaty from Romney’s disclosures, but come to different conclusions on whether that qualified as a failure to disclose. Vanity Fair offers support for Gibbs’s contention: “Romney failed to list this entity on several financial disclosures, even though such a closely held entity would not qualify as an ‘excepted investment fund’ that would not need to be on his disclosure forms.” But AP disagrees: “The omissions were permitted by state and federal authorities overseeing Romney’s ethics filings, and he has never been cited for failing to disclose information about his money.”

This is an excellent opportunity for a real fact check. Two creditable sources reached diametrically opposite conclusions. Whether Romney failed or not hinges on whether the Vanity Fair or AP analysis of disclosure rules is correct. Conducting a fact check that settled the disagreement would be a great way to determine whether or not there was a failure. PolitiFact, however, punted:

There is some question whether he was required to disclose this asset. In 2003, on the eve of running for governor, he put Sankaty into a blind trust. Thereafter, he disclosed the presence of the trust and might not have been required to say what was in it. It is also unclear whether the nominal value of Sankaty would be high enough to warrant disclosure.

Despite their inability to determine what Romney was or was not required to disclose, PolitiFact assigned a “True” rating to Gibbs’s statement because Romney “did not disclose it.” However, PolitiFact never documented the disclosure requirements needed to support Gibbs’s actual words that Romney “failed to disclose that.”

Until it can be determined whether the absence of Sankaty on the disclosures was proper or not, Romney’s failure, or lack thereof, is in doubt. PolitiFact’s failure is, however, doubtless. When you present yourself as a fact checker, you need to check the facts.

Holy Misrepresentation, Batman!

There has been recent discussion about the homophonic relationship between Mitt Romney’s company, Bain, and the villain of The Dark Knight Rises, Bane. Recent musings on that subject by Rush Limbaugh earned a “Pants on Fire” rating from PolitiFact. They concluded:

Limbaugh suggested it’s no accident that in a movie coming out four months before the presidential election, the villain bears the same name as the company formerly run by Romney and now being attacked by Democrats.

In politics, conspiracies are everywhere if you look hard enough. But Limbaugh’s superpowers of persuasion can’t make this theory stand up. The villainous Bane first appeared in Batman comic books in 1993, long before Romney entered presidential politics. Even the character’s creator called a suggested link “ridiculous.” We rate this statement Pants on Fire!

With other media members making claims similar to PolitiFact’s, Limbaugh responded: “I never said that the villain was created by the comic book character creator to be part of the 2012 campaign. I said that Democrats were going to use it, which they are.”

The transcript PolitiFact used as a source offers more support for the radio host than it does the fact checkers. Only two sentences discuss whether the name Bane was selected because of Romney. The first is a question: “Do you think that it is accidental that the name of the really vicious fire breathing four eyed whatever it is villain in this movie is named Bane?” After asking that question, Limbaugh was interrupted and went off on a tangent. When he returned to the subject, he continued: “So, anyway, this evil villain in the new Batman movie is named Bane.  And there’s now a discussion out there as to whether or not this is purposeful and whether or not it will influence voters.” After noting it, Limbaugh never takes a side in the discussion of purposefulness. The remainder of his monologue relates to the ways the presidential campaigns can and will use Bane and other Batman related themes.

The PolitiFact article implies that Limbaugh used his “superpowers of persuasion” to advance the idea that the villain Bane was named in reference to Bain Capital. The transcript shows that he didn’t even try to make that argument.

The Contributor

PolitiFact recently published an article reviewing a statement that Maryland Governor Martin O’Malley made about Mitt Romney’s term as governor of Massachusetts:

When it comes to what little bit of job creation was happening in Massachusetts, it was happening in the public sector at six times the rate that it was happening in the private sector.

PolitiFact rated the statement as “Half True” because the growth O’Malley was talking about referred to the increase in state government employees, not all public sector employees. It seems a little harsh to allow Romney to take the full blame for this. The state legislature did override 700 of the 800-plus items he vetoed. Although this could be fleshed out more fully, it isn’t worth the time. Something else very strange and interesting appeared in the same article.

The following was the contribution of the Brookings Institution’s Gary Burtless to this PolitiFact endeavor:

“The private sector accounted for a bigger share of all payroll employment gains in Massachusetts during Gov. Romney’s term,” said Gary Burtless, a labor market expert with the centrist Brookings Institution. (As we’ve noted before, Burtless contributed $750 to Obama’s campaign in 2011. However, in 2008 he provided advice on aspects of labor policy to the presidential campaign of John McCain, and he has worked as a government economist and served on federal advisory panels under presidents of both parties.)

He added: “The more legitimate criticism of Gov. Romney’s job creation record is that total payroll employment growth lagged far behind the experience of the rest of the country.” For 2002 to 2006, the national growth rate was 5.2 percent, while Massachusetts’ rate was just .7 percent.

That second paragraph is a stunner. It offers no information about the veracity of O’Malley’s statement, but it appears nonetheless. Burtless uses a statistic Democrats have been using of late to attack Romney. But that statistic is useless in analyzing the matter at hand.

It’s very odd that PolitiFact chose to quote a non-sequiturial Democratic talking point from an Obama contributor. The oddness only increases when they themselves clarify the talking point in the next sentence. Perhaps this is some kind of quid pro quo for consenting to a PolitiFact interview. It will be interesting to see if other interviewees will be their given 25 words of editorial space in future articles.

Post-normal fact checking

Glenn Kessler, the Washington Post’s Fact Checker, decided to review the following statement made by Mitt Romney when he appeared on Fox News:

“We should be seeing numbers in the 500,000 jobs created per month. This is way, way, way off from what should happen in a normal recovery.”

Kessler decided that the best way to “check” this was determine how frequently 500,000 jobs were created in a month in the last 65 year. The figure of 65 years was chosen because that is “Mitt Romney’s age.” Despite the unlikelihood of finding a viable counter-argument to such rigorous logic, let us endeavor.

Apparently in only nine of the last 784 months (that’s 65 years worth of months) did job growth exceed 500,000 jobs. This was the foundation that Kessler used to label Romney’s statement “a stretch.”  So what number wouldn’t have been a stretch?

First, we need to use Kessler’s methodology to find the typical month — the month where half the other months had a lower job gain, and half a higher job gain. There are folks much smarter than me who call this the median. The median monthly job growth for the last 65 years is 146,000 jobs. That’s the Kesslerian normal.

This may meet the Kessler criterion for normal, but Paul Krugman calls it ‘inadequate.’ That’s actually generous because Krugman called a gain 200,000 jobs per month ‘inadequate.’ Kessler’s magic number of 146,000 per month doesn’t even rise to that level. As Krugman points out, “we need at least 100,000 jobs a month just to keep up with working-age population growth.” The extra 46,000 jobs a month offer little in the way of growth, especially during a recession.

In light of the serious issues with the methodology, it is quite difficult to consider such analysis a “fact check.”