Sunday, November 27, 2011

PolitiFlub: Scott Walker vs. the unions

Just a reminder:  I exercise selection bias in deciding which PolitiFact items to critique.  By no means do I read all of them.  I look at ones I think are likely to contain errors based on their subject matter.  I tend not to have any difficulty finding errors, and I expect that if I looked at more of the items I'd find many more errors.

I don't have time for do a detailed critique of every PolitiFact item in which I find flaws.  But sometimes I have to point out a significant mistake even if I don't do the full critique.  That's what "PolitiFlub" items are for.

This item concerns a spectacular offense against the principle of charitable interpretation, committed by PolitiFact Wisconsin.  PolitiFact took issue with Walker's description of his clash with Wisconsin's public sector unions:
He focused sharply on labor unions, which fought legislation by Walker and Republican lawmakers to curtail collective bargaining and force  public workers to contribute more toward pensions and health care. That push, which became law, attracted massive and prolonged protests in Madison.

"I asked the unions to pay into their own health care insurance (just as their Wisconsin neighbors do) and they said I was being unreasonable," Walker’s letter said. "I requested that they contribute toward their own pensions (just as their Wisconsin neighbors do) and they screamed it was unfair."
What was the problem?

In PolitiFact's eyes, Walker made it look like he was asking the unions to agree to the changes in pension and health care contributions:
(T)he portrayal of "asking" the unions rewrites history, leaving the misleading impression there was give and take with labor.
Why is this a PolitiFlub?  Isn't PolitiFact obviously correct?

Shame on you if you think so.

PolitiFact conveniently overlooks that "ask" and "request" and their permutations are routinely used where the associated behavior represents a demand.

Teacher:
"I ask that you all take your seats."

Politician:
"With this tax hike we're simply asking the filthy stinkin' rich to pay their fair share."

Still think I'm kidding?
5. to demand; expect: What price are they asking? A little silence is all I ask.
And, yes, the word "request" works the same way:
The principle of charitable interpretation must receive evenhanded application in fact checking to ensure fairness.  Every subject, without exception, is entitled to charitable interpretation.  Skipping the step according to whim or bias leads to the straw man fallacy and other errors.

Friday, November 25, 2011

Another embarrassing juxtaposition for PolitiFact

Oy vey.

A little over two weeks ago, I panned a PolitiFact Florida fact check that found "Mostly False" the claim that one can register to vote without proving citizenship.

Just now I stumbled over a parallel fact check from PolitiFact Wisconsin.  Apparently state Rep. Donna Seidel claimed that a new Republican-backed law would allow persons without firearms training to receive a state permit to carry a concealed weapon.

In the Florida case, a person registering to vote needs to sign an affidavit affirming possession of legal citizenship status.  That's it.  That's the proof requirement unless one registers through the "motor voter" program.

In Wisconsin, the law continues to require firearm training but does not specify the amount of training required.  Here's how PolitiFact Wisconsin summarized it:
States typically do one or more of the following: certify the training organization, their instructors or their courses; mandate specific topics for training; set a minimum number of hours; require the instructor to sign the training certificate.

Wisconsin now requires none of those.
The finding?  They ruled Seidel's statement "True" without reservation.  The justification for the ruling provides an amazing comparison with the reasoning applied by PolitiFact Florida with respect to voter registration:

PolitiFact Wisconsin:
The law is still on the books requiring an instructor-led training course, but it’s hard to prove that someone skipped it. So the door is open to "untrained" individuals getting a permit.
PolitiFact Florida:
Individual supervisors of elections do have the ability to determine if an applicant is an U.S. citizen, Cate said, though it would be difficult for a supervisor to ascertain citizenship.
In the former case the difficulty of proving the lack of qualification serves as a key point supporting the Democrat's statement.  In the latter case, the difficulty of proving the citizenship of a registered voter is apparently completely ignored as a reason contributing to the truth of the claim.

If a Wisconsin Democrat raises concerns about untrained persons carrying firearms, PolitiFact sees a problem with a soft standard for proving training status.

If a Florida Republican raises concerns about illegal immigrants registering to vote, PolitiFact finds hardly any problem with a comparable standard of proof.

Despite the similarity of the claims and the related set of facts, the Democrat receives a "True" while the Republican receives a "Mostly False" from the Truth-O-Meter.

And doesn't that fit pretty well with the liberal preferences for easy voting and stricter controls on firearms?

The "PolitiFact heuristic"

A thoughtful college student wrote a pretty good blog post about the usefulness of PolitiFact ratings.

Peter William Hurford overall does a handy job of pointing out the limited conclusions we can legitimately draw from "Truth-O-Meter" ratings, but ends up offering PolitiFact a bit too much credit.

The problem stems from an early assumption:
Enter PolitiFact. PolitiFact is a website located at politifact.com that aims to “fact-check statements by members of Congress, the White House, lobbyists and interest groups”. Run by the St. Petersburg Times, an independent newspaper, PolitiFact elaborates on their history and promises “that no one is behind the scenes telling us what to write for someone else’s benefit. We are an independent, nonpartisan news organization. We are not beholden to any government, political party or corporate interest. We are proud to be able to say that we are independent journalists. And for that, we thank Nelson Poynter.”

Thus we probably have enough information to establish PolitiFact as reasonably trustworthy, and a sufficiently reliable source of information that we can draw upon it to approximate knowledge about the trustworthiness of candidates in a heuristic.
PolitiFact reliable?

Credit Hurford with using the language of caution ("probably" "sufficiently reliable"), but his premise about PolitiFact's reliability is a judgment call resting on little more than a presumption of PolitiFact's competence and fairness.

An abundance of anecdotal evidence shows PolitiFact failing journalistic standards as well as its own standards while strongly suggesting that PolitiFact exhibits the leftward ideological tilt we might expect from the typical aggregation of journalists.  Cases like one I wrote about hours ago, where two very similar fact checks essentially gave a pass to the lone Democratic Party entity involved would doubtless give a reasonable person like Hurford pause if he was aware of their existence.

I should emphasize again that Hurford's reasoning is mostly solid.  He does a good job of pointing out the types of conclusions the findings at PolitiFact would fail to reasonably support.  But at the bottom line, the "PolitiFact heuristic" is probably less useful than Hurford's earlier example of heuristic reasoning:  trusting elite opinions.

And an informed voter may well be more likely than the writers and editors at PolitiFact at identifying statements that discredit a candidate.



Thursday, November 24, 2011

Another shifting standard at PolitiFact

Remember the terrible job PolitiFact did fact checking the Florida Democratic Party's claims about Republican gubernatorial candidate Bill McCollum's supposed vote for his own pay raises?

No?  Well, I'm not so surprised.

PolitiFact Florida revisited the issue yesterday and did a decent job of explaining the pay raise process that made the claim about McCollum so dubious.  Not to worry, Florida Democratic Party.  I'm sure your "True" rating is permanently ensconced in PolitiFact's Hall of Records.  And it's far too late to erase any political damage your ad did to McCollum that year.

This time it's Republican against Republican, with Republican senatorial candidate George LeMieux making a similar claim about Connie Mack's voting record.  LeMieux received a "Half True" rating.

LeMieux is a dirty lyin' scoundrel.  The Florida Democratic Party is pure as the driven snow.

Time for me to write another futile letter to PolitiFact ...
Dear Amy Sherman (cc Aaron Sharockman),

Thanks for doing a credible job with your story on congressional pay raises.  I sincerely appreciate seeing you do such a nice job on it.

On the downside, your inconsistency at PolitiFact just cracks me up.  Another PolitiFact team did a parallel fact check of a very similar claim about Bill McCollum back in 2010 and found it true without reservation.  That claim was made by the Florida Democratic Party, for what that's worth.

Will you dare demand a revision of that item, shedding on it some of the light from the superior reporting this time around?  It's a bit late to help out McCollum, of course, whose reputation took the damage from PolitiFact's earlier fact check, but it might ease a conscience or two.
 Ball in PolitiFact's court.

Wednesday, November 23, 2011

Fatally flawed design

An MIT grad student has designed software that would highlight suspicious sentences as a type of fact-checking help.

But there's a problem:
Schultz is careful to clarify: His software is not designed to determine lies from truth on its own. That remains primarily the province of real humans. The software is being designed to detect words and phrases that show up in PolitiFact’s database, relying on PolitiFact’s researchers for the truth-telling. “It’s not just deciding what’s bullshit. It’s deciding what has been judged,” he said. “In other words, it’s picking out things that somebody identified as being potentially dubious.”
So the software is next to useless.

Hasn't Schultz heard of Annenberg Fact Check?  Still biased, but not the joke that is PolitiFact.

Tuesday, November 22, 2011

Grading PolitiFact: Mitt Romney's Obama ad and talking about the economy (Updated x2))

Context matters -- We examine the claim in the full context, the comments made before and after it, the question that prompted it, and the point the person was trying to make.
--Principles of PolitiFact and the Truth-O-Meter
PolitiFact delivered on the above principle this time--after a fashion.


The issue

On occasion I'll provide an expanded clip of PolitiFact's visual presentation of the story to help emphasize the way our supposed fact checkers mislead the audience.

This is another of those times.

(clipped from PolitiFact.com)
PolitiFact sends the message that it is beyond merely false that Barack Obama said "If we keep talking about the economy, we're going to lose."


The fact checkers:

Louis Jacobson:  writer, researcher
Bill Adair:  editor


Analysis:

This is an amazingly inept effort by PolitiFact.

You've been warned.

PolitiFact:
On the eve of a presidential trip to New Hampshire on Nov. 22, 2011, Mitt Romney’s campaign released an ad targeting President Barack Obama. In the ad, the Romney campaign used a quote that prompted an immediate counterattack from the Obama camp, which argued that it had been taken out of context.
Wait a minute.  The Obama camp complained about the quotation being taken out of context?  So Obama actually said what Romney claimed he said?   What about the headline and deck material making it look the opposite?  What about that "Pants on Fire" rating when PolitiFact's rating system until recently had a category for accurate statements that take things out of context ("Half True")?

What's going on here?  Jacobson had better come through with a spectacular explanation for this one.

PolitiFact:
The 60-second ad, called "Believe in America," is designed to contrast "candidate Obama from 2008 with President Obama of today," highlighting "his failures in between," according to the Romney campaign.

The ad contrasts a 2008 campaign speech by Obama with text on the screen that criticizes Obama’s economic record, including, "Greatest Jobs Crisis Since Great Depression," "Record Home Foreclosures" and "Record National Debt."

The ad then has a clip of Obama saying, "If we keep talking about the economy, we’re going to lose."
PolitiFact fails to make clear that every one of Obama's words come from that same 2008 campaign speech. And the Romney ad prefaces the quotations with the year in which they were made:  2008.

Here's the ad:



The ad contains nothing to cue the viewer that Obama was speaking about the 2012 election in the last clip.  Paying attention to the context, the viewer is left to figure out what Obama was talking about in 2008.  Obviously Obama isn't talking about his own re-election prospects on Oct. 16, 2008--not in those words.  Obama hadn't been elected at that point. The election didn't take place until November of that year.

PolitiFact somehow fails to see it:
The clear implication is that Obama believes that his economic record is so bad that he will lose in 2012 unless he can steer the conversation away from the economy.
Why would Obama have any beliefs at all about his economic record as president way back in 2008?  This supposed "clear implication" occurs only if the viewer either ignores the context or has an unaccountably difficult time taking obvious clues from the context.  We're not in Sherlock Holmes territory, here.

But PolitiFact rolls with it:
But the Obama camp, among others, immediately charged that the clip was taken out of context. Was it?

Here’s what Obama said in the October 2008 speech, which came about two weeks before he defeated Sen. John McCain:

"Even as we face the most serious economic crisis of our time, even as you are worried about keeping your jobs or paying your bills or staying in your homes, my opponent's campaign announced earlier this month that they want to ‘turn the page’ on the discussion about our economy so they can spend the final weeks of this election attacking me instead," Obama said in the speech. "Sen. McCain's campaign actually said, and I quote, ‘If we keep talking about the economy, we're going to lose.’"

So the comment is drastically different than the way it's portrayed in the Romney ad. Obama was actually saying that his opponent’s campaign three years earlier had said, "If we keep talking about the economy, we're going to lose." That context is not included in the Romney ad -- and leaving it out sends a profoundly different message.
1)  The claim from the Obama camp that the quotation was taken out of context deserves its own fact check. 

2)  PolitiFact is correct that Obama was referring to something said by the McCain campaign.

3)  PolitiFact is incorrect that omitting the context sends a profoundly different message.   I will illustrate.

Let's make it a 35 second ad instead of a 30 second ad, adding in the full quotation from Obama, and let PolitiFact explain it just like before:

The ad contrasts a 2008 campaign speech by Obama with text on the screen that criticizes Obama’s economic record, including, "Greatest Jobs Crisis Since Great Depression," "Record Home Foreclosures" and "Record National Debt."

The ad then has a clip of Obama saying, "Sen. McCain's campaign actually said, and I quote, ‘If we keep talking about the economy, we're going to lose.’"

What is the message of the ad with the context included?

No doubt someone could complain that in our augmented version Romney makes it look like McCain is saying that Obama can't get re-elected on his economic record.  But we can even make it a 45 second ad, including context sufficient to ensure that every viewer understands that McCain was talking about the economy in 2008 and the GOP prospects for the presidential election and it still doesn't change the point of the ad.

The point of the ad is that if it wasn't proper to run on the economy in 2008 then it's even worse to run on the economy in 2012, with many economic indicators far worse than they were in 2008.

And the fact that Obama mockingly brought up what the McCain campaign said pretty well confirms that Obama did not think the 2008 economy was good election material for the incumbent party.  Is Obama supposed to think otherwise in 2011-2012 with unemployment over 8 percent, the housing market still a shambles and an anemic growth rate?

That's the point of the ad.  It doesn't change with the context added.  And the fact that it doesn't change with the context added means that the quotation was not taken misleadingly out of context. 

PolitiFact, of course, fails to see it that way:
Our ruling

We certainly think it’s fair for Romney to attack Obama for his response to the economy. And the Romney camp can argue that Obama’s situation in 2011 is ironic considering the comments he made in 2008. But those points could have been made without distorting Obama’s words, which have been taken out of context in a ridiculously misleading way. We rate the Romney ad’s portrayal of Obama’s 2008 comments Pants on Fire.
Kudos to PolitiFact for at least admitting the real point of the ad is legitimate despite missing the real point.  Only viewers completely unable to appreciate the significance of Obama's remarks from 2008 could miss the real point of the ad.  The real point flies whether McCain said it of himself, whether McCain said it of Obama or whether Obama said it of himself.

In a situation like that no additional context is needed.  The Romney campaign was justified in omitting it.


The grades:

Louis Jacobson:  F
Bill Adair:  F

What can you say about a team that wrote a knee-jerk response to the Romney ad?  What can you say about a team that produced a presentation that makes it appear false that Obama said something that Obama did say?  What can you say about a team that corrupted the Truth-O-Meter's supposed grading system in favor of its subjective knee-jerk response?

They're journalists reporting badly.



Update:

ABC News serves up a handy reminder that PolitiFact isn't the only impossibly inept news source out there.  ABC ran the following under the headline "Mitt Romney ad misquotes President Obama":
Mitt Romney’s inaugural TV ad of the 2012 campaign aired today in New Hampshire just as President Obama traveled to the state, but the ad immediately came under fire from Democrats and fact-checkers for incorrectly quoting Obama.

The White House, the Democratic National Committee and Obama’s re-election campaign accused the Romney campaign today of unfairly twisting the president’s words.
We have people reporting the news who do not know what is and what is not a misquote.

It's stuff like this that makes it plain to so many, regardless of a want of an easy scientific proof, that the leftward tilt of media ideology results in left-tilted news reporting.


Update 2:

Jim Nolte at Big Journalism evaluated the Romney ad the same way I did, and he published first:
Watch Romney’s ad again. The point wasn’t “look at what Obama said!” The point was that the statement about talking about the economy is true when it comes to Obama.  You could add the full context and it might even hit Obama harder because of the obvious irony. Moreover, campaigns do this kind of thing all the time.
Word.

But don't expect PolitiFact to pay any attention to the criticism unless it is picked up and amplified by its primarily liberal fan base.

Monday, November 21, 2011

Hope 'n' change at PolitiFact (Updated x2)

I keep hoping that criticism will influence positive change at PolitiFact, the fact checking arm of the St. Petersburg Times (soon changing its name to the Tampa Bay Times).

Well, a positive change occurred at PolitiFact recently.

Unfortunately, it was of the "one step forward, two steps back" variety.

For some time I've carped about PolitiFact's inconsistent standards, and in particular its publishing of two different standards for its "Half True" position on the "Truth-O-Meter."

The recent change probably stemmed from a message I sent to an editor at the paper's city desk (sent Nov. 9):
PolitiFact has created a problem for itself through inconsistency.  During the site's earlier years a page called "About PolitiFact" gave information about how the "Flip-O-Meter" and the "Truth-O-Meter" supposedly operate.  The page includes a description of each of the "Truth-O-Meter" rating categories.

More recently, editor Bill Adair posted an item called "Principles of PolitiFact and the Truth-O-Meter."  The problem?  The definition for "Half True" is different than the one PolitiFact posted for well over a year prior.  Compounding the problem, PolitiFact has kept both versions online through now.

1)  The statement is accurate but leaves out important details or takes things out of context.
2)  The statement is partially accurate but leaves out important details or takes things out of context.

I'll be interested to see the eventual remedy.  Which items over PolitiFact's history went by which definition? Was a change made in Feb. 2011 or before without any announcement?  How can PolitiFact legitimately offer report cards and "Truth Index" ratings if the grading system isn't consistent?  Those are questions I'd imagine readers would have if they realized PolitiFact is using two different definitions for the same rating.  I don't expect you to answer them for my sake (not that I would mind if you did). 

Good luck to all sorting this one out.
The eventual remedy is apparently to simply change the longstanding definition at "About PolitiFact" to match the newer one at "Principles of PolitiFact and the Truth-O-Meter" without any fanfare--indeed, without any apparent notice whatsoever.  I detect no admission of error at all and no acknowledgment that PolitiFact changed its standard.

The move seems consistent with the desire of the mainstream press to avoid doing things that "undermine the ability of readers, viewers or listeners to believe what they print or broadcast."

Sadly, I'm not at all surprised.

On the positive side, the definitions are now consistent with one another.

On the negative side, PolitiFact either created a past illusion where Truth-O-Meter ratings used the old system or else created a fresh illusion that past ratings follow the new system.  And went about it in about the least transparent way possible.


Update:

Good luck to PolitiFact retroactively changing the dozens (perhaps hundreds) of places on the Web that republished the original definition of "Half True."


(Clipped from PolitiFact.com; click image for enlarged view)

Contact PolitiFact Wisconsin.  They didn't get the memo yet.  And PolitiFact Texas has the same problem.


It's not the crime, it's the coverup.


Update 2:


It's also worth remembering PolitiFact's agonizing decision to change "Barely True" to "Mostly False."

"It is a change we don't make lightly," wrote Bill Adair.

How do you like that?  A change in the wording of a rating gets a reader survey prior to the change and an article announcing the change.  A change in the definition of a rating--a much more substantial change--gets the swept-under-the-rug treatment.

Sunday, November 20, 2011

Video: Blue Angels with Dixie Dregs soundtrack

What's not to like?   I can't resist posting this.



Is Steve Morse slightly stoked at having his music used for this video? Morse is, after all, a pilot.

Saturday, November 19, 2011

Video: "The European Fiscal Crisis and Lessons for America"

Yes, it's time for a flirtation with "rule 5" blogging, featuring yet another economics instruction video from the Center for Freedom and Prosperity.



I like the line about the "safety net" turning into a hammock.

Hat tip to Hot Air.

Friday, November 18, 2011

PolitiFlub: Who gambles the most?

Argh.  Perhaps Media Matters is better than PolitiFact after all.

The latest improbable flub from PolitiFact involves the question of gambling demographics.

PolitiFact examines a statement supposedly from John Stemberger.  Here's the headline blurb:

The largest number of gamblers are "from the poorest segments of the population."  

John Stemberger on Thursday, October 20th, 2011 in a website
However, it turns out that the portion in quotes was in turn quoted by Stemberger from another source (bold emphasis added):
The blog post from Oct. 20, 2011, uses partial quotes from Wayne Grudem, a theology professor at Phoenix Seminary in Arizona and author of the book Politics According to the Bible:

"My own judgment is that large commercial gambling outlets such as casinos and state-sponsored lotteries bring much more harm to a society than the benefits they generate (such as tax revenue)… First, it is socially harmful (and fiscally regressive) because the largest numbers of gamblers comes from the poorest segments of the population. Second, (it) leads to an addiction to gambling … and this addiction destroys marriages, families … and increases societal breakdown. Third, studies have shown that where gambling businesses are established, crime rates increase."
Every bit of the information PolitiFact is checking, as a matter of fact, comes from quotations of Grudem.

But there's more.

PolitiFact is good enough to link to the page in Grudem's book from which the material ultimately came.  But PolitiFact does not provide what may be a key part of the context of Grudem's claim, and ignores that potential key element in its reasoning.

Grudem:
(S)erious objections can be brought against gambling, or at least commercial gambling as a business.  A number of studies have shown that gambling brings negative effects in a society, and these must be seriously considered.  First, it is socially harmful because the largest number of gamblers comes from the poorest segments of the population, who make unwise decisions and trap themselves deeper and deeper in debt. Second, the existence of gambling businesses leads to an addiction to gambling on the part of a certain percentage of the population, and this addiction destroys marriages, families, and any hope for career advancement ...
Note that Grudem is sourcing his objections to an unnamed "number of studies."  So, if we keep each of the successive statements in context, the fact check ought to be about whether those studies say what Grudem claims they say (compare Jon Stewart).

Yet the fallout lands on John Stemberger without even apparently considering the studies to which Grudem alludes.  Though it's worth noting that Grudem apparently does not cite the studies specifically, fact checkers thereby obtain no excuse for ignoring the original context of the claim.

As bad as the failing makes PolitiFact appear, it gets even worse.

Note the wording of the claim from Grudem:  "the largest number of gamblers comes from the poorest segments of the population."  Though Grudem refers to segments plural, PolitiFact proceeds to interpret it in the singular.  That interpretation leads to what may represent a straw man version of Grudem's argument.

Note the response elicited from one of PolitiFact's expert sources:
"In my opinion this is a poorly worded and misleading statement," said David Just, an economics professor at Cornell, who has studied poverty and lotteries. "By no means does this group constitute the majority of those playing the lottery. Those in poverty are just 16 percent of the U.S. population."
It's possible that Grudem had a poverty-level demographic in mind.  But is an assumption warranted?

PolitiFact simply fails to handle the material objectively.  Stemberger did not make the statement attributed to him other than by quoting his source.  Grudem deserves the fact check if anyone, and PolitiFact takes his statement out of context and applies to it something less than the charitable interpretation.


Afters:

I'm still shaking my head over the opening paragraph:
In the ongoing war about gambling in Florida, some critics have turned to the Bible to state their case.
Some critics may have turned to the Bible to make their case, but it doesn't have much to do with this fact check, which concerns a criticism of the lottery based on economics and harm to society.

The paragraph doesn't belong in this story.  Good grief.

Wednesday, November 16, 2011

More D'OhbamaCare

The story "Another ObamaCare Glitch" by Jonathan H. Adler and Michael F. Cannon, appearing in the Wall Street Journal, identifies yet another problematic aspect of President Obama's signature piece of legislation:
Even if ObamaCare survives Supreme Court scrutiny next spring, its trials will be far from over. That's because the law has a major glitch that threatens its basic functioning. It's so problematic, in fact, that the Obama administration is now brazenly trying to rewrite the law without involving Congress.
It seems the PPACA requires state-run insurance exchanges to provide premium assistance but contains no provision for the federal version where states elect not to organize an exchange.

Adler and Cannon provide plenty of details about the can of worms thus opened, so read the whole article.

Sunday, November 13, 2011

The EPA should regulate airplanes

While researching a PolitiFact fact check, it occurred to me that the Environmental Protection Agency ought to regulate airplanes.  Not the emissions from airplane engines, but the airplanes themselves.

This notion stems from the definition of air pollution that EPA and the courts have used to justify shoehorning carbon dioxide into the regulated pollutant category (bold emphasis added):
While greenhouse gases and their impacts have been a matter of concern for years, these gases were not definitively determined to be an air pollutant covered by the CAA until the Supreme Court resolved that issue affirmatively in 2007. In addressing this issue, the Court looked to the definition of “air pollutant,” which is defined as “any air pollution agent or combination of such agents ... which is emitted into or otherwise enters the ambient air.” Massachusetts v. EPA, 549 U.S. 497, 506 (2007) (“Massachusetts”).
Airplanes enter the ambient air with every takeoff, unless we're supposing the mere manufacture of a plane causes it to enter the air (albeit grounded).  It seems like the statutory text would foreclose any exclusion of airplanes from the EPA's regulatory sphere.  Certainly planes regularly cause human endangerment via crashes and skydiving accidents (not to mention the heightened risks to human life associated with warplanes!), so it makes complete sense for the EPA to require permits of plane owners.

Just sayin'.

Tuesday, November 08, 2011

Grading PolitiFact (Florida): Alan Hays, proof of citizenship and voting

Words matter -- We pay close attention to the specific wording of a claim. Is it a precise statement? Does it contain mitigating words or phrases?
--Principles of PolitiFact and the Truth-O-Meter
If anything about an anecdote can strongly suggest an ideological bias, it is the complete ease with which PolitiFact can ignore its supposed principles.


The issue:

(clipped from PolitiFact.com)


The fact checkers:

Amy Sherman:  writer, researcher
Aaron Sharockman:  editor


Analysis:

Words matter.  Here are the words Alan Hays used during the committee meeting, via the PolitiFact story (bold emphasis added):
"Before we design a district anywhere in the state of Florida for Hispanic voters, we need to ascertain that they are citizens of the United States. We all know there are many Hispanic speaking people in Florida that are not legal, and I just don't think that it's right that we try to draw a district that encompasses people that really have no business voting anyhow. If we know registered voters are people who have proven their citizenship then that's a completely different story, but I'm not aware of any proof of citizenship necessary before you register to vote."
Hays says he isn't aware of any proof of citizenship necessary in Florida, before one can register to vote.

Watch the PolitiFact twist:
Hays said you don't need to prove that you're a citizen in order to cast a ballot.
Words matter?

If words matter then we should expect PolitiFact to note the difference between saying one does not know of a requirement and saying that no requirement exists.  If PolitiFact subjected its own claim to the "Truth-O-Meter"--and the thing actually worked like it's supposed to--I don't see how it could rate higher than "Mostly False."

It seems that words matter at PolitiFact's discretion.  That is, according to subjective criteria.

Fact checking a dubious paraphrase rather than the subject's actual words already constitutes justification for failing grades, but even apart from this blunder the fact check takes an astonishingly errant course.

Is it necessary to prove citizenship in order to register to vote in Florida?

PolitiFact takes up the question by initially examining the motor voter registration system, perhaps reasoning that the emphasis is appropriate given the popularity of that mode of registration:
Floridians can register to vote a variety of ways -- but the most common registration method is at driver's license offices. In 2010, more than 270,000 people registered to vote at driver's license offices. The second-most common method was what the state division of elections calls the "other" category, which includes registering in person at supervisors of elections offices.
Motor voter registration apparently does require proof of citizenship:
In Florida when drivers apply for a driver's license, they must provide proof that they are in the country legally. U.S. citizens could show a passport or a proper birth certificate to verify their citizenship. Immigrants who are not U.S. citizens would have to provide the proper visa.
But for non-citizens who wish to vote, the motor voter identification requirements represent no real barrier:
But what if someone walks into a supervisor of elections office and asks to register to vote?

In those cases, people registering to vote must sign an oath on a registration application attesting that they are qualified to vote and that all the information on the application is true. (Usually at driver's licenses offices applicants who want to apply to register to vote don't have to sign that particular form since the office already has their signature as part of the driver's license application process.)

The application includes a question: Are you a citizen of the United States? If you answer no, the form says "you cannot register to vote."

We asked state Division of Elections spokesman Chris Cate if the state does anything to verify citizenship.

"The answer is no," he said. "The law doesn't require someone to provide proof of citizenship when they register. If they swear, attest and sign under oath that their information is accurate and that they are a citizen we will accept their voter registration."
Voter registration through an elections office, then, basically utilizes the honor system.  If you say you're a citizen then you can register to vote regardless of whether you couldn't provide proof of citizenship when you received your driver's license.

Game over?  It may look like it, but the story continues with a description of the torments waiting in Hell (felony charges, anyway) for non-citizens who lie on their registration form.

It isn't relevant.  Whether or not the state metes out punishments to those it isn't even seeking who have improperly registered, they're allowed to register without proving citizenship.

Then there's this puzzling statement from PolitiFact:
Individual supervisors of elections do have the ability to determine if an applicant is an U.S. citizen, Cate said, though it would be difficult for a supervisor to ascertain citizenship.
Huh?  I'd like to see that unpacked.  Supervisors of elections have the authority to verify citizenship but verifying citizenship is difficult?  That's my best guess at translating the meaning.  And it isn't very reassuring in that form.

But PolitiFact still isn't finished.

It turns out that election experts Ion Sancho (FEC filings here) and Emogene Stegall say there's no problem with non-citizens voting:
Sancho said the idea that illegal residents are voting is "laughable it's so wrong."

"We are not seeing any problem with illegal citizens voting in the U.S. anywhere, not just in Tallahassee or Florida. It's a canard that illegal individuals are registering and voting. ... Voting requires putting your name and address on an official document and that is not something undocumented individuals tend to do."

The local elections supervisor in Hays' own district, Lake County, sees it the same way.

"We've never had a problem with illegal voting in Lake County, no way,'' said Emogene W. Stegall, who has served in the county's election's office for 40 years.
PolitiFact interviews "impartial experts."  Or Democrats.  Apparently either will do.

Opinions from Sancho and Stegall are not necessarily wrong simply because they are liberals.  Simply observe that their opinions are accompanied by the very thinnest of evidences, the best of it being Sancho's claim that undocumented individuals tend not to put names and addresses on official documents.

Seriously, what's the harm in putting a name and address on voter registration if elections supervisors do not attempt to verify citizenship?  Do these two have any data in support of their expert opinions?

What does the Truth-O-Meter conclude?
Our ruling

Sen. Hays said that Florida doesn't require "any proof of citizenship necessary before you register to vote." There is a kernel of truth here: According to the state Division of Elections, state law doesn't require new voter applicants to prove their citizenship in some physical sense. They simply must sign a sworn statement attesting that they meet the voting requirements -- including being a citizen.
Again setting aside the, ah, liberal interpretation of Hays' statement, why is there simply a "kernel of truth" to the proposition that it isn't necessary to prove citizenship in order to register to vote in Florida?  PolitiFact subsequently mentions caveats, so let's look there for our answer:
First, willfully lying on a voter registration form about your citizenship status can lead to a felony conviction.
Failing to bring your vehicle to a complete stop at a stop sign can lead to a traffic citation.  Enforcement is really the key, isn't it?  The state admits that it has no active enforcement mechanism attached to the voter registration system other than the easily circumvented identification requirements associated with the motor voter program.
Second, the most common way to register to vote in Florida is during the process of obtaining a driver's license -- in 2010, 57 percent of those who registered to vote in Florida chose that method. And during that process, people are asked to verify their citizenship.
Is this a joke?

If registration through the driver's license program isn't a necessary requirement to register to vote then the identify checks associated with that program are not required in order to register to vote.  It's basic logic.  And it is already admitted that one can go straight from the driver's license office to the supervisor of elections and register on the honor system.
Lastly, we have to consider the experiences of two actual supervisor of elections, who said there is no widespread problem of illegal immigrants registering to vote.
We have no evidence that either of the helpful liberal supervisors of elections base their claims on empirical data.
For those reasons, we rate this claim Mostly False.
 Those were rationalizations, not reasons.

The fact check attributed a claim to Hays that Hays did not make and then used a ridiculous set of excuses to obscure the truth of the made-up claim.

I'm sputtering trying to think of condemnatory language equal to the task of lambasting PolitiFact over this fact check.


The grades

Amy Sherman:  F
Aaron Sharockman:  F

Both members of the team could have earned failing grades simply for making up the claim attributed to Hays.  But this duo when beyond the pale by ignoring clear evidence that the claim they graded "Mostly False" was essentially true.

We have yet another instance of a PolitiFact team earning the label "journalists reporting badly."

Monday, November 07, 2011

The St. Petersburg Times and Charlie Crist's share of public campaign dollars (Updated)

I don't have the St. Petersburg Times to kick around any longer.

The newspaper has changed its name to the Tampa Bay Times to (allegedly) more accurately reflect its distribution pattern.  Dollars to donuts the name change is at least partially designed to increase the paper's appeal to those in the areas surrounding St. Petersburg.

Regardless, I'm holding the Times responsible for what the Times did in the past, namely ballooning in its reporting the amount of money former Florida governor Charlie Crist received through public campaign financing.

As reported in earlier blog posts, I notified the Times of its error back in June 2010, shortly after the flawed stories were published.  I eventually went all the way up the chain of command, to the point of sending an email to the Times' president.

None of that had any effect, apart from the author of the news report (Steve Bousquet) apparently agreeing that the report on Crist's share of the campaign funds was incorrect.

Still eager for a greater sense of how the Times feels about the charge of inaccuracy, this month I started phoning the paper's city desk.  Editor Jim Booth also appears to agree that the paper's reporting was in error, and he told the he has forwarded the problem to those who are supposed to deal with it.

Booth sounds unhappy about the mistake, for what it's worth.  And he was unerringly cordial on the phone.

So now the waiting game is on again.

We'll see what happens.


Update:

W00t!

Yesterday the Times mostly fixed both errors.  Complaining by phone must be magical or something.

I can quibble with the correction notice:
Correction: During his 2006 run for governor, Charlie Crist received $3.3 million in public financing. A previous version of this story included a different figure.

[Last modified: Nov 07, 2011 01:52 PM]
It's not like the previous version said "$3.5 million."  I'd prefer greater transparency.

Sunday, November 06, 2011

Grading PolitiFact: Mitt Romney and $95 billion in savings

Context matters -- We examine the claim in the full context, the comments made before and after it, the question that prompted it, and the point the person was trying to make.
--Principles of PolitiFact and the Truth-O-Meter

Ordinarily I do not participate in PolitiFact's uber-lame "Lie of the Year" voting, but I'll consider the above as a write-in entry when the time comes.


This issue:

(clipped from PolitiFact.com)


The fact checkers:

Angie Drobnic Holan:  writer, researcher
Bill Adair:  editor


Analysis:

Even though I see stuff like this from PolitiFact pretty much every week I still can't believe my eyes sometimes.

PolitiFact rates "False" Republican presidential candidate Mitt Romney's claim that repealing ObamaCare would save $95 billion.  So let's see how PolitiFact reasons it out (bold emphasis added):
Mitt Romney has recently been emphasizing one of the favorite themes of the tea party movement: cutting government spending.

He’s been getting pretty specific about some of his ideas, both in an op-ed in USA Today and in speeches on the campaign trail. His plans include ending subsidies for Amtrak, stopping funding the Corporation for Public Broadcasting, and eliminating foreign aid to countries "that oppose America’s interests."

He also wants to save money by rolling back President Barack Obama’s signature health care law.

See the change?  PolitiFact avers that Romney's op-ed talks about ways to cut government spending.  That's in the first graph.  By the third graph, the idea has morphed into saving money.  The two can mean the same thing but do not necessarily mean the same thing.

PolitiFact:
He made the point in the USA Today op-ed, suggesting he would  "repeal ObamaCare, which would save $95 billion in 2016."

We were surprised by his suggestion. As we remembered the health care negotiations, Democrats took pains to make sure the 2010 health care law was projected to reduce the deficit, and they bragged repeatedly about their numbers.
PolitiFact's surprise is apparently a product of a liberal bias.

Romney was writing about reducing spending in that section of his op-ed, not about the net effect of repealing the health care reform bill.  It's easy to prove (bold emphasis added):
There are three ways to reduce spending, which combined, will achieve a fiscal turnaround of this size.

First, eliminate every government program that is not absolutely essential. There are many things government does that we may like but that we do not need. The test should be this: "Is this program so critical that it is worth borrowing money to pay for it?" The federal government should stop doing things we don't need or can't afford. For example:
 
•Repeal ObamaCare, which would save $95 billion in 2016.
Taking context into account, Romney's list of bullet points are his suggestions for cutting spending.  It is perfectly legitimate to talk about spending cuts as distinct from overall deficit reduction.  Somehow this is lost on the fact finders at PolitiFact.  Even setting aside the misleading nature of the CBO scoring for the PPACA (the doc fix and CLASS, to name two obvious examples), PolitiFact constructs a straw man version of Romney's argument.

As a result, the subsequent paragraphs lauding the supposed budget savings from the PPACA are irrelevant.  Romney was talking about cutting spending, not cutting the deficit as such.

PolitiFact sums up:
So according to the CBO analysis, a full repeal of the bill would reduce the deficit by $16 billion in 2016, much less than the number Romney cited.
Romney cited a number for a reduction in spending.  PolitiFact grades him on deficit reduction.  Of the two, PolitiFact is the one engaged in partisan spin.

It's disgusting to label this schlock as a fact check.

The judgment from On High:
If Romney had only criticized the law as an expansion of government spending, he would have been on firmer ground. Instead, he asserted that a repeal of the law would save significant money -- $95 billion. In fact, the law included new taxes and cost reductions so that the actual savings for the year he cited would be much smaller -- $16 billion. And, over the long haul, repealing the law actually adds significantly to the deficit. So we rate his statement False.
Romney did criticize the law in terms of its expansion of government spending.  And not spending $95 million saves $95 million in spending.


The grades:

Angie Drobnic Holan:  F
Bill Adair:  F

Together, they are journalists reporting badly.



Friday, November 04, 2011

AP Spin Meter flunks logic, spins story

The Associated Press (unwittingly?) turned "Spin Meter" into a pun with today's story on a supposed inconsistency by the GOP.

The AP story, carrying Donna Cassata's byline, charges that Republicans regard defense spending as a job creator while also claiming that government spending does not create jobs.

Cassata's story contains four main problems.

First, Cassata produces very little quoted material to support her claims about what GOP politicians supposedly say.

Second, the quotations she uses fail to support the story's claim of a reversal, inconsistency or contradiction (however one wishes to interpret her claim):
Running for re-election, Rep. Howard "Buck" McKeon, R-Calif., said in February 2010 that the stimulus package did not create new jobs. In a statement about the economy and jobs now on his website, McKeon says "congressional Democrats and the administration continue to insist that we can spend our way out of this recession and create jobs, but the numbers just don't add up."

But at a hearing last week, McKeon, now the committee chairman, argued against cuts to the military, saying, "We don't spend money on defense to create jobs. But defense cuts are certainly a path to job loss, especially among our high-skilled workforces. There is no private sector alternative to compensate for the government's investment."

He later added, "While cuts to the military might reduce federal spending, they harm national security and they definitely don't lead to job growth."
There simply isn't any contradiction in McKeon's statements.

Cassata appears to simply assume the presence of a contradiction, and tries to sustain the impression for the reader's sake with stuff like this:
Asked about the competing statements, a spokesman for McKeon, Claude Chafin, said they were "not inconsistent" because the defense industry is a unique recipient of federal dollars.

Note how it is taken for granted that the statements compete with one another.  Pulling the two word snippet ("not inconsistent") from Chafin makes his response look like a thin denial if not a case of the fallacy of special pleading.

One should note that special pleading is not a fallacy where special circumstances obtain, not that McKeon ultimately needs to argue that point.

Third, and stemming from the second failure, Cassata fails to make any type of logical case for McKeon (or the GOP) offering an inconsistent view of government spending.  As McKeon made very clear in his statement, concerns over jobs lost from government cuts are separate from concerns over net job creation.

Fourth, Cassata enters the dreaded realm of liberal media bias.  By the time her story concludes it could pass for an op-ed extolling the virtues of government spending as a means of producing jobs, with that thesis reliant on the writings of a Keynesian economist with a strong record of giving money to liberal politicians.

I will emphasize that my point is not that Robert Pollin's work should not be trusted because of his partisanship.  The point is that a fact check should not arbitrarily take sides in a controversy and call it a fact check.  The truth is that Pollin's opinion has no real relevance to McKeon's consistency.  It may have relevance to McKeon's position that stimulus spending did not work, but if that's the case then Pollin should have to explain the apparently dismal results of the stimulus bill.

Conclusion:  The Spin Meter is spinning.