You are not logged in. LOG IN NOW >

Fact-Checking the Candidates: PolitiFact

BY Joshua Levy | Monday, August 27 2007

Whenever I watch a debate or a candidate's YouTube video or I read a mass email making claims about budgets being raised, policies getting enacted, or support being offered, I'm always wondering, "is anyone independently tracking whether or not any of this is true?" 

The folks at the St. Petersburg Times and Congressional Quarterly, led by journalist Matt Waite, have stepped in to help. They've teamed up to produce PolitiFact, a new site that is tracking the claims made by the candidates and registering them on the "Truth-O-Meter."  It's not much of surprise that few statements are ever "True"; instead, most fall somewhere around "Mostly True," "Half True," and "Barely True."  I love the idea that something is just barely true...

This isn't the first site to tackle the problem of half-truths in politics; the Annenberg Center's has been taking a broad approach to "holding politicians accountable" already.  But PolitiFact (say that five times fast) is the first to focus solely on the presidential election.

Waite, an investigative journalist with the St. Petersburg Times, is running the project with very little help; he even developed the web site himself, the first time he'd done so.  "The site is a part of democracy," he told me.  "The whole thing with the Internet is that the ink is never dry."  Despite limitations on staff and money, and more ideas than he could develop by himself, Waite decided to produce a basic version of the site.  "We're talking a lot about PolitiFact 2.0," Waite says.  "We're not done developing this thing at all.  We're learning every minute this thing is online."  Matt, welcome to the iterative world of the web!

So far, the site's data is compiled by journalists and researchers who are paid to pore through a database created by Waite to find distortions, embellishments, or outright lies uttered by the candidates.  There are no gotcha moments per se; instead the half-truths (or barely-truths) most of us suspect we are hearing are laid bare.  YouTube videos serve as evidence, and sources are prominently displayed and linked to.  You can search by candidate, and there's even an "Attack File" that specifically unpacks attacks made by the candidates on their opponents. 

A video on the front page shows Chris Dodd's campaign claiming that "Barack Obama and Hillary Clinton have changed their positions (on the Iraq war withdrawal) to follow Chris Dodd."  Only half-true, says PolitiFact.  "Dodd is correct that his fellow candidates backed an amendment he co-sponsored... But Dodd exaggerates his role... There’s no evidence that Dodd had any influence on Obama and Clinton."

This is great stuff, and I think will become a valuable resource.  But while the St. Petersburg Times should be commended for creating a site that is web native (rather than just reposting print material), one element is conspicuously lacking: the crowd.

"The whole site is inspired by Adrian Holovaty’s manifesto on the fundamental way newspaper websites need to change," Waite wrote on his blog
Adrian’s main theme was that certain kinds of newspaper content have consistent pieces that could be better served to the reader from a database instead of a newspaper story. I built PolitiFact with that in mind... We’re experimenting with greater transparency, by listing our sources for each statement and story. We’re taking into account that YouTube is going to be a significant factor in this election by embedding YouTube videos of statements when we can find them.

This is all good, but truly taking advantage of the read/write web doesn't just mean embedding YouTube videos and being more transparent (though those are both big steps).  It also means opening up to your readers, and letting them in on the process.   In this case, it could be more cost-effective to open up the database to the readers of the site, and let them pore through the records and make the connections.  As Jay Rosen's recent overtures to crowdsourcing and the various projects of the Sunlight Foundation have shown us, the crowd is smarter than a few individuals; open up the process to them and you'll be amazed at what they find.  Waite says he's open to this and other ideas; the problem is minimal resources at his disposal.  With luck, the project will take off and we can look forward to a new, participatory PolitiFact 2.0.