Is Truth-O-Meter the real issue in Maddow’s latest blast at PolitiFact?

The Tampa Bay Times’ fact-checking site PolitiFact has drawn another heated rebuke from MSNBC anchor Rachel Maddow, who accuses it of “ruining fact checking” and being “truly terrible.”

But at the risk of looking like a homer — the Times signs my checks as its media critic — I think Maddow’s gripe with PolitiFact boils down to the same thing that’s rankled other critics: the site’s Truth-O-Meter rulings. (Additional disclaimer: Poynter owns the Tampa Bay Times.)

On Tuesday, Maddow took issue with PolitiFact ruling as “Half True” a statement from tennis legend Martina Navratilova that “in 29 states in this country you can still get fired for not just being gay but if your employer thinks you are gay.” That number is the amount of states with no statewide law banning employment discrimination for sexual orientation.

But PolitiFact noted that several factors work against making blanket statements based on a lack of state laws. Some government employees have protections against sexual-orientation discrimination even in those 29 states. Cities in states lacking such laws have passed their own legislation banning workplace discrimination based on sexual orientation. (Philadelphia and Pittsburgh are two examples.) Some employers have union rules and written internal policies barring such discrimination. And some laws banning gender discrimination can also protect gay people, depending on how a case is argued.

But are such exceptions enough to make Navratilova’s statement “Half True”?

I’m betting that’s what bothered some who read the PolitiFact analysis. I would have given Navratilova’s words a rating of “Mostly True,” since a) PolitiFact didn’t seem to calculate how many people might be protected by these exceptions; and b) the exceptions seem like minor ones. As I see it, “Half True” overstates the case because it implies a substantial error or falsity.

The Truth-O-Meter, which assigns statements to six categories on a scale from “True” to “Pants on Fire” for out-and-out falsehoods, has been both PolitiFact’s most successful and most controversial element.

On the one hand, it provides a handy, quick method for branding PolitiFact, recognizing its rulings and communicating its decisions. Anyone looking to laud or blast a statement can use this shorthand; Daily Show host Jon Stewart even used his smartphone to read former GOP candidate Herman Cain a “Pants on Fire” ruling during the program’s visit to the Republican National Convention in Tampa.

But on the other hand, the Truth-O-Meter can provide an easy source of criticism. Maddow also blew up at PolitiFact in February 2012 when the site ruled “Mostly True” a claim by Republican Sen. Marco Rubio that “Americans are majority conservative,” citing a 2011 Gallup poll that found 40 percent of Americans identified as conservative, compared to 21 percent liberal and 35 percent moderate.

Again, this is ruling I would dispute, because 40 percent is a long way from 51 percent. I probably would have ruled it “Mostly False,” because the real number isn’t a majority, even though it is the largest category of the three measured by the poll. (After taking a lot of criticism, PolitiFact eventually changed its ruling to “Half True.”)

PolitiFact’s explanations of its rulings are an effort to go beyond the literal truth of a fact or set of facts to judge the overall impact of a statement. In politics, it is easy to lay out three true statements and reach a false conclusion; the subjective Truth-o-Meter rulings are a way of addressing this issue. And by laying out the facts it weighed in reaching a ruling, PolitiFact lets the reader make his or her own decision. As long as the facts PolitiFact presents in its arguments are true, criticism that the site is “ruining fact-checking” overlooks much of what it does.

Some critics have asked whether PolitiFact has set out to tweak liberal sensibilities with some of its rulings, perhaps offering a harsher Truth-O-Meter setting to look even-handed in political squabbles. People who work on the site insist that isn’t happening, but readers can look over PolitiFact’s rulings and decide for themselves.

That’s an important difference between PolitiFact and Maddow’s latest critique of it. Even while lambasting PolitiFact for a supposed error, Maddow never fairly explained the facts assembled by the site to challenge Navratilova’s statement, dismissing them as “unrelated information.” And that makes it tougher for Maddow’s viewers to judge if her analysis was fair.

So in this case, it seems, both sides might have a little to learn about fair arguments and rulings.

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.

  • CarlosMu

    The politifact ranking was so absurd that you wonder what’s up with them. Why did they choose this item in the first place, a basic fact uttered by an athlete on TV? It looks like they went out of their way to discredit it.

    And by the way, great kudos to Maddow for covering this. Her tone is pretty much perfect for Politifact’s absurdity. Bewilderment, disbelief, frustration, etc. Talk about “truth”, watch Maddow’s segment on this.

  • http://twitter.com/InisMagrath Inis Magrath

    Caveats about “certain” government employees, or “some” private companies with their own policies — or what ever other issues may be in play — NONE of these things render untrue Navratilova’s statement that “in 29 states in this country you can still get
    fired for not just being gay but if your employer thinks you are gay.”

    I guarantee that PolitiFact would rule “half true” at best, if not “mostly false,” the statement that cigarette smoking causes cancer. After all, not EVERYBODY who smokes gets cancer, right?

  • Fromens1987

    Though not as deeply biased as say Fox news, I’d have to say yes it’s
    predictably right of center. Consistently, I’m not so sure.
    Inconsistently consistent, if there is such a thing.

  • http://twitter.com/ZebraFactCheck Bryan W. White

    Alfred Ingram wrote:

    “If a statement from the “left” is mostly true, but labeled half true,
    and a statement from the right is a half truth, but is labeled mostly
    true, we are describing a right of center bias.”

    What examples did you have in mind? Or is it a pure hypothetical?

    Ingram continued:

    “Though not as deeply biased as say Fox news, I’d have to say yes it’s predictably right of center.”

    If you don’t have but one example, how is that an example of bias? Isn’t it a good idea to count up how many times PF miscalculated its ratings before reaching a conclusion that the example(s) constitute bias? And how will you know that you did a better job calculating the rating than PolitiFact?

    I’d suggest that using PolitiFact’s grading system to second-guess PolitiFact is a bit of a mistake. The “Truth-O-Meter” system stinks. Deggans is right to call it subjective (I’d qualify that as saying it’s significantly subjective rather than wholly subjective). If you correct a subjective rating with another subjective rating the end result is a subjective judgment, isn’t it?

  • Alfred Ingram

    If a statement from the “left” is mostly true, but labeled half true, and a statement from the right is a half truth, but is labeled mostly true, we are describing a right of center bias. The real question is, “Is this bias consistent and predictable?” Though not as deeply biased as say Fox news, I’d have to say yes it’s predictably right of center. Consistently, I’m not so sure. Inconsistently consistent, if there is such a thing.

  • http://twitter.com/ZebraFactCheck Bryan W. White

    The “Truth-O-Meter” is just one of PolitiFact’s many problems (I’m delighted to see you describe it as “subjective”–does that belong in a news story?)

    PolitiFact just doesn’t do fact checking well, and the Rubio example you gave serves to illustrate. If Rubio had said most Americans self-identify as conservative, then PolitiFact did its fact check the right way. But that’s not what Rubio did. He said he knows most Americans are conservative because they run from the “liberal” label. You gave the self-identification level for “liberal” at 21 percent. If somebody wants to fact check Rubio they should look his clearly-stated inference and consider the context. Is it reasonable to count a person as conservative if they reject the “liberal” label?

    In context, is Rubio just giving an opinion? Probably. One could argue that point. But it’s pretty clear (or ought to be) that PolitiFact went about the fact check the wrong way. They built a straw man version of Rubio’s argument. The straw man version probably only warranted a “Half True” at best. But Rubio’s real point was probably that Republicans can win elections nationally by appealing to things the electorate dislikes about liberalism. And that’s not a bad point. It’s better than “Half True.”