PolitiFact walks back second ruling in a month, this time on Rubio claim

PolitiFact has revised its piece from Feb. 14, which found U.S. Sen. Marco Rubio’s contention that “the majority of Americans are conservatives” to be “Mostly True.” It’s now “Half True.”

It’s the second time in about a month that PolitiFact, a project of Poynter’s Tampa Bay Times and a 2009 Pulitzer winner, has revised a ruling after an outcry. In January, the fact-checking site gave President Obama a “Half True” for his claim on jobs growth, then upgraded it to “Mostly True” after “hearing lots of feedback from readers.”

In an editor’s note on the revised Rubio post, PolitiFact says it heard from lots of readers steamed by the original ruling, which hinged on reasoning I’ve read four times this morning and still don’t quite get: “Rubio said that the majority of Americans are conservative. A respected ongoing poll from Gallup shows that conservatives are the largest ideological group, but they don’t cross the 50 percent threshold. So we rate his statement Mostly True.”

The editor’s note says that “the debate centered on whether to judge Rubio on his literal statement or the underlying point. We try to balance that question in many of our rulings.”

In the new post, PolitiFact adds a little more chiaroscuro to its ruling: When pollsters offer voters the option to identify themselves as moderates, conservatives never account for more than a plurality. Take away the “moderate” option, as in a Politico-George Washington University poll, PolitiFact says, and many more Americans identify themselves as conservatives.

The original decision, which didn’t mention the Politico-GWU poll, turned PolitiFact into a punching bag on Rachel Maddow’s MSNBC show and many liberal blogs. In a forensic piece on the larger Maddow-PolitiFact contretemps, Washington Post media blogger Erik Wemple argued that PolitiFact’s “Truth-O-Meter” ruling obscured the subtleties behind its argument: “Viewers and the rest of us appreciate the end product, nothing more. On this front, PolitiFact erred most egregiously on the Marco Rubio contention.”

In Politico, Dylan Byers has argued that PolitiFact should drop the “Truth-o-Meter,” saying it engenders tension between simple rulings and nuanced arguments. PolitiFact head honcho Bill Adair told Byers the meter rulings are “a key part of PolitiFact’s work.”

Adair tells me by email that the site doesn’t track how many times it’s revised rulings, but he estimates that it’s happened about 10 times for the national site and perhaps 10 to 15 for the state sites.

Earlier: Krugman, liberals argue Politifact’s Lie of the Year is not a lie (Poynter) | Do fact-checking operations merely confirm liberal bias? (Poynter) | PolitiFact draws ire for checking ‘Glee’ (Poynter)

We have made it easy to comment on posts, however we require civility and encourage full names to that end (first initial, last name is OK). Please read our guidelines here before commenting.

  • http://www.facebook.com/profile.php?id=504633504 Dan Mitchell

     I just want to say that this is the second time today I’ve seen “refute” used correctly, and it’s making me happy. Though I’m sure tomorrow some NPR reporter will use it to mean “disagree with” and I’ll be deflated all over again.

  • http://www.facebook.com/profile.php?id=519877869 Al Cross

    This shows the fundamental problem with Politifact: its reliance on pigeonhole categories (yes, “Pants on Fire” is funny). As someone who fact-checked political ads for part of his living for several years, I have always objected to the Truth-o-Meter. It’s a silly gimmick that puts marketing ahead of journalism. It’s why I rarely cite PolitiFact and prefer FactCheck.org.

  • Anonymous

    Byers is right. The “Truth-O-Meter” is a marketing gimmick which detracts from the actual fact-checking done by PolitiFact. 

    “The “Truth-O-Meter” allows PolitiFact to market its research — which is painstaking and time-consuming — to a political discourse that doesn’t have time to read its analysis.” 

    The sad thing is PolitiFact’s research is usually pretty good; it’s their “Truth-O-Meter” rating that gets them in trouble. (I find Maddow’s takedowns of PolitiFact particularly humorous in this regard since she uses PolitiFact’s own research to refute their own ratings.) 

    Context isn’t quantifiable; either get rid of the “Truth-O-Meter” or stick with empirical facts. 

  • http://www.facebook.com/people/Bryan-White/100000501966225 Bryan White

    The changed ruling is incomprehensible as a matter of logic.

    The new version states that the changed ruling is “bolstered” by new reporting.  That’s false.  The new ruling of “Half True” is easier to justify from the old reporting.

    Key points:  The new reporting added a poll of likely voters that supported Rubio more than anything included in the original version of the story. Also in the new story, PolitiFact quoted an expert source (Lake) as saying that forcing moderates to express a preference for liberal or conservative produced a superior result for predicting voting behavior (providing a reason to prefer the poll of likely voters to the two others).

    Based on the evidence, this walk back was a cave to external pressure rather than a reconsideration of the evidence.