Sunday, June 01, 2008

The PolitiFact problem (Updated)

I've taken the PolitiFact fact-checkers to task a number of times over the past months. Despite the backing of Congressional Quarterly and the St. Petersburg Times, PolitiFact's efforts place it in the lower echelon in the fact-checking biz, well below Annenberg and the Washington Post. Instead, PolitiFact rates on a par with Media Matters, albeit with a lower degree of blatant partisanship than the latter.

I've decided to ramp up my coverage of PolitiFact since nobody else is doing it. In this post I'll provide an overview of PolitiFact and a review of its problems.

clipped from

PolitiFact is a project of the St. Petersburg Times and Congressional Quarterly to help you find the truth in the presidential campaign. Every day, reporters and researchers from the Times and CQ will analyze the candidates' speeches, TV ads and interviews and determine whether the claims are accurate. >> More


Though PolitiFact is expressly intended to "help you find the truth in the presidential campaign," their presentations tend to focus on small issues in a big way, and present judgments in a manner that lends itself to distortion.

An example from the recent past shows the emphasis on the unimportant. PolitiFact composed an entry to address the following statement from Barack Obama: "Right now, an employer has more of a chance of getting hit by lightning than be prosecuted for hiring an undocumented worker. That has to change."

Obvious hyperbole, right? PolitiFact jumped all over it and found Obama had spoken false. The chance of being hit by lightning is less than that for prosecution for hiring an undocumented worker. The big surprise is that PolitiFact did not break down lightning strikes geographically and according to personal habits (like habitually carrying around a big metal antenna during a thunderstorm) in order to evaluate the claim.

And here is the thing: Though the self-description at PolitiFact claims that its writers and researchers review speeches and ads and use their own news judgment to choose topics, I received a response from Matthew Waite ("Developer of PolitiFact") that indicated that news judgment was influenced by reader questions. So, I guess I can judge PolitiFact "half true" regarding the way they describe their methods. Seriously, allowing reader input to steer the process introduces a potential bias in choosing the subject matter. In this case, PolitiFact compounded the bias problem by somehow forgetting to even mention the concept of hyperbole in its condemnation of Obama's truth-telling.

Another recent case shows another problem with PolitiFact. Check out the first four sources listed for this item about an Obama attack on McCain:

Sources: YouTube, "Obama Criticizes McCain on Lobbyists" from May 21 speech in Tampa

New York Times, "Top Aide to McCain Defends Lobbying Work" by Michael Luo, May 19, 2008

MSNBC, "Adviser Calls Lobbyist Scrutiny ‘Nonsense’" by Carrie Dann, May, 19, 2008

Media Matters, "CNN, First Read repeated McCain adviser’s false claims about campaign manager’s lobbying history" on May 20, 2008

See the sore thumb? That's right, "Media Matters" listed as though it was not partisan news commentary but was instead a trusted news source. Serious researchers would not list Media Matters among their list of sources even if they were naive/biased enough to use it for research in the first place. Media Matters will not have contributed any reporting on the facts of the story. Just slanted analysis. And speaking of the potential bias in having readers help choose the subject matter in political fact checking ... Media Matters is a pretty good bet to involve itself in activism on behalf of its editorial position.

In addition to outsourcing its news judgment and using partisan sources, PolitiFact makes another key error that will forever keep credibility beyond its feeble grasp: the use of the cheesy "Truth O Meter" graphic. For some reason that utterly escapes me, they're proud of it:
PolitiFact (pronounced puh-lit’-eh-fact) is bolder than previous journalistic fact-checking efforts because we’ll make a call, declaring whether a claim is True, Mostly True, Half True, Barely True or False. We even have a special category for the most ridiculous claims that we call “Pants on Fire.”
"Bolder." That's supposed to be good.

The heart of PolitiFact is the Truth-O-Meter, which we use to rate the candidates’ claims and attacks.

The Truth-O-Meter is based on the concept that – especially in politics - truth is not black and white. Depending on how much information a candidate provides, a statement can be half true or barely true without being false.

Begging to differ, truth is black and white. However, it's not always proper to append a label because one statement could combine enough black and white together so that the statement looks gray from a distance. In replying to Matthew Waite, I reiterated my past observation that the "Truth O Meter" is ultimately misleading, particularly when applied to statements such as the Obama hyperbole mentioned above. PolitiFact rated Obama's statement "False." But to a reasonable person, Obama's point was that prosecution for hiring illegals is very rare--and no more than that. In particular, Obama was almost certainly not making any literal comparative claim between prosecutions and lightning strikes.

Yet PolitiFact found the statement false!

If the folks at PolitiFact insist on grading hyperbole and other figures of speech false, then the grading system they chose is inadequate. It would be fair to rate hyperbole false while at the same time pointing out how the less-than-true literal meaning actually accords with a literally true message from the candidate: Prosecution for hiring undocumented workers is too rare.

Unless PolitiFact comes up with a more precise way of grading, the "Truth O Meter" should be tossed on the scrap heap. I know the Times is currently under the control of graphic designers and not dedicated journalists, but the designer of the graphic will get over it. Eventually.

So, until PolitiFact gets serious about its work by hiring me or somebody like me to cure the problems before they are presented to the world via the Internet, I'll be putting increased emphasis on exposing the more spectacular gaffes.


My periodic polemics addressing PolitiFact flim-flam are well underway.

While I anticipate that PolitiFact's mistakes exhibit a pattern of liberal partisanship, my critiques are not designed to establish any such pattern. I will critique what suits me, and ordinarily that will tend to involve my support of conservative issues that tend to receive unfair treatment by the media.

I do not think that journalists ordinarily do anything other than attempt to report and fact-check objectively (though the current political season has challenged my belief on that point). Journalists are simply subject to some of the same misconceptions and misperceptions as ordinary folks and it creeps into their writing as a result.


  1. Media Matters may be biased sometimes, but asserting that it can't be used as a source shows your innate bias,and undermines your credibility in assessing politifact.

    M Matters usually documents what they are saying with factual support, in full context usually, and where it came from. You can judge bias usually right right by what they offer. If they took out the apparent bias, they could make many of the same points and be one of the more reliable sources of bias and information on the web. but to claim that citing them if what was cited was accurate and documented, is itself not credible, is a bit ridiculous

    Here's bias for ya concerning McCain.

  2. Administrator,

    You're partly correct, in that using Media Matters as a reference does not automatically indicate bias, and I have already acknowledged in a more recent post:
    "Since my previous general assessment of PolitiFact, written back in 2008, one distinct improvement seems to have occurred. PolitiFact no longer lists the left-leaning fact-check site Media Matters on its list of sources. Though I can imagine ways that Media Matters could be used without sacrificing objectivity, the vast majority of fact-checks should not require any material at all from that source."

    The fact-check I was using served as a good example. There was no good reason to list Media Matters as a source. Unless relevant material was produced at Media Matters or it served as the host for material not readily available elsewhere.

    It bolsters my argument that PolitiFact went for most of a year without again using Media Matters as a listed source.

    Thanks for stopping by and contributing your opinion.


Please remain on topic and keep coarse language to an absolute minimum. Comments in a language other than English will be assumed off topic.