October 12, 2010

@PolitiFactWisc does it again

This thing has gotten ridiculous, as if it hadn't already.

PolitiFactWisc purports to evaluate this claim of Senator Russ Feingold's:
"According to independent analysis, unfair trade deals have resulted in the loss of over 64,000 jobs in Wisconsin."
First, the PolitiFactWisc team independently audits Feingold's source for the figure, and finds it ... dead nuts accurate.

PolitiFactWisc then questions the independence of the source, and finds its independence lacking because its compilers support Senator Feingold's stance on free trade deals (which doesn't make it any less independent or accurate) so the PolitiFactWisc team checks with an alternative source that supports the free trade deals in question.

That source confirms:
Gary Clyde Hufbauer, a Peterson Institute expert, said the data [Feingold uses] is commonly used by researchers as the best available for showing how trade can close factories.
Emphasis added.

Indeed, asserts the PolitiFactWisc team, 64,000 "may vastly understate the number of jobs lost" due to the free trade pacts.

Vastly: enormously. Immensely. Colossally.

Finally, says PolitiFactWisc, "those numbers leave out the other side of the equation — jobs created through the deals."

Yet PolitiFactWisc doesn't cite a single source — independent or otherwise — for any "jobs created through the deals." And for all of this, PolitiFactWisc determines Feingold's claim, "Half True."

Even though the PolitiFactWisc team hasn't the foggiest notion of whether those "jobs created through the deals" would simply be offset by the "vast understatement" of jobs lost represented by the 64,000 figure, and doesn't even bother trying to find out.

So why, exactly, is Senator Feingold's claim "Half True"?
Wisconsin lost over 81,958 manufacturing jobs (or 15.6 percent) during the NAFTA-WTO period (1993-2009), according the Bureau of Labor Statistics. This figure is for TOTAL manufacturing employment, so takes into account both jobs created by exports and jobs displaced by imports, among other factors.
Because it's too conservative?

We (generously) rate PolitiFactWisc's latest entry, "Half Assed."*

Moreover, as both Jay Bullock and Mike Plaisted have correctly pointed out, PolitiFactWisc has lately taken to "evaluating" the claims of various (Republican) candidates that nobody had even challenged.

What on Earth is the point of that?

* We also fault Feingold for using "over" instead of "more than."

Earlier: A Milwaukee paper's Bizarro World PolitiFact®


John Foust said...

No place to comment on that web page. Again the old media doesn't quite get thenew media. Heavens, if they let just anyone perform their own analysis, they might be shown wrong.

gnarlytrombone said...

Curious about your meta-diagnosis, IT. The incompetence and/or guile of the reporters on this beat is manifest. But do you think the format is hopelessly arbitrary?

illusory tenant said...

For one thing, this particular item is devoted in part to assessing what constitutes "fairness," which in this case is a subjective determination, not a fact.

Of course there's nothing objectionable about an attempt to break down the constituent facts that might contribute to one or the other party's notion of fairness -- like, for example, an appeals court applying a series of tests to evaluate whether due process obtained throughout a trial -- but simply because there exists some nebulous dispute over what's fair doesn't justify designating the figures provided by Feingold (and verified by the newspaper) as only "half true" (whatever that even means).

Otherwise don't call the project Politi-"Fact."

gnarlytrombone said...

I like that analogy to legal tests. It sure would help to know what standards of reasoning are being applied.

There's been scattered backchatter in recent years on these themes - bird-dogging political rhetoric in addition to "facts" - in journalism circles. And from what I've read the founders of the original PolitiFact wrestled with them.

Communication experts who've weighed in stress that the focus should be on the structure of argument, not on policing vocabulary. The latter results in just as you say: a morass of subjective values. The former means reporting on the deployment of specific, misleading tropes and tactics which, say the experts, is reportable as fact. That is to say, it's possible to assess how a claim is used in addition to checking the claim's veracity.

Such reporting only works if 1) the journalist understands how to analyze an argument or appeals to someone who does and 2) shows their work; i.e., clearly explicates why the tactic is misleading. As you've ably noted, applying a true/false label in such a case often leads to bizarre contradictions with the analysis.

The Politifacters are failing badly at the rhetoric beat on both counts, if this indeed what they're attempting. But I'm less and less sure what exactly the major malfunction is locally, especially after that piece on Walker and his salary. That was simply a big, sloppy kiss.

illusory tenant said...

Great comment, thanks. Tomorrow we're going to find out whether Sean Duffy really was "a no show."

John Foust said...

Uh oh. Professor Shh-n-Shhis asking for more compassionate criticism of his reading comprehension.

illusory tenant said...

What the heck's that post got to do with this one?

gnarlytrombone said...

He's rehearsing lines, obviously. He's gonna dress up as Judge Posner for Halloween.