Well, this is very meta, but seems a useful tool. Real Clear Politics analyzes the methods of a number of fact-checking organizations such as Politifact and Snopes.
But who shall fact-check the fact-checker checkers themselves? My faith in RCP is not exactly unquestioning.
Facebook and others who call for fact-checking everything someone posts online seem to be looking for an absolute authority, and there isn't one. The only result of initiatives like this is an Office of the Censor. (And I think they realize this. Authority is good, as long as it's Amenable Authority.)
From a brief look-through, RCP seems pretty open about their methods and gives their data, so pretty much anyone could check the fact-checker checker.
Also, lots of interesting stuff that you could aggregate and use to do some possibly fruitful analysis. E.g., the names of the people who did the checking, whether the item checked was a fact or opinion, the source of the claims, the verdicts, the sources used to check, etc.
On 'source of claim,' it also shows the most common sources for claims. For example, Politifact, the NYT, the WaPo, and FactCheck.org all check Trump a lot more than any other source of claims. Snopes and the Weekly Standard, on the other hand, check claims on social media most (Facebook, Twitter, etc.).
I like the RCP tool because it applies a fairly transparent calculation to sources. It doesn't have to draw any conclusions; it just gives us some neutral facts with which we might draw conclusions. That puts it way ahead of most fact-checking operations, in my view.
I do use Snopes. It's not bad at catching silly web themes that are based on year-old news that's breathlessly announced on Facebook with the subtitle, "Why haven't we heard about this on the TV news this week?" It's completely unreliable when it comes to debunking a politically charged claim. It will sometimes grudgingly admit some literal truth to a political claim it doesn't like, but then includes a lot of political argument of its own before concluding "partially true." If it stuck to facts I'd find it more credible.
But who shall fact-check the fact-checker checkers themselves? My faith in RCP is not exactly unquestioning.
ReplyDeleteFacebook and others who call for fact-checking everything someone posts online seem to be looking for an absolute authority, and there isn't one. The only result of initiatives like this is an Office of the Censor. (And I think they realize this. Authority is good, as long as it's Amenable Authority.)
From a brief look-through, RCP seems pretty open about their methods and gives their data, so pretty much anyone could check the fact-checker checker.
ReplyDeleteAlso, lots of interesting stuff that you could aggregate and use to do some possibly fruitful analysis. E.g., the names of the people who did the checking, whether the item checked was a fact or opinion, the source of the claims, the verdicts, the sources used to check, etc.
On 'source of claim,' it also shows the most common sources for claims. For example, Politifact, the NYT, the WaPo, and FactCheck.org all check Trump a lot more than any other source of claims. Snopes and the Weekly Standard, on the other hand, check claims on social media most (Facebook, Twitter, etc.).
Interesting stuff.
I like the RCP tool because it applies a fairly transparent calculation to sources. It doesn't have to draw any conclusions; it just gives us some neutral facts with which we might draw conclusions. That puts it way ahead of most fact-checking operations, in my view.
ReplyDeleteI do use Snopes. It's not bad at catching silly web themes that are based on year-old news that's breathlessly announced on Facebook with the subtitle, "Why haven't we heard about this on the TV news this week?" It's completely unreliable when it comes to debunking a politically charged claim. It will sometimes grudgingly admit some literal truth to a political claim it doesn't like, but then includes a lot of political argument of its own before concluding "partially true." If it stuck to facts I'd find it more credible.