Thursday, April 1, 2010

Oh, Nate, You Know I Love You, But We Have to Talk

I would like to hope to like to think that my rivalry with Nate Silver at could be three things that it might today be something closer to zero of, at least the way I've conducted myself so far. I'd like to hope to like to think, specifically, that it could be respectful, jealousy-free, and constructive -- at least to the extent that it might elevate the quality of the work of at least one of us. Indeed it would appear that some of my blog commentators have missed the frequency with which I've said complimentary things about Mr. Silver and his site -- something I've done often. He deserves them.

But the current post on his site, penned under his by-line, is, well, just nutty, frankly.

The article, written by a guy whose credential as a political analyst was secured for all-time by using an elaborate and econometrically rigorous model to correctly predict forty-nine of fifty states in the 2008 Presidential election, argues that there is no proof of the widely savored rumor that the extreme right wing in this country is hurting its own cause by refusing to fill out its census forms. To prove this, Mr. Silver posts the current national average for already-returned forms at the top of a data-table -- about 50%, so far -- and then lists the return rates of all twenty-one states that have voted Republican in the three most recent Presidential elections. Some of these states, as he notes, have higher census return-rates than the national average; others have lower ones.

And then? Then, Mr. Silver -- who, let's face it, is self-evidently a better econometrician with a hangover, a migraine, and a concussion than I'll be on the best day of my life -- does a very, very, very curious thing: he averages the return-rates of all twenty-one of those states, with positively no regard to differences in their populations. And this, folks, this is supposed to be the smoking-gun proof that the red-state-refusal hypothesis is unfounded. To consider just how sloppy a statement this is on his part, one needs to know just which states are being over-represented in his average and which ones are being under-represented, and by how much. The July 2009 estimate for the population of Texas, for example, is 24,782,302. At a 44% response-rate to the census mailing, Texas is lagging 6% behind the national average -- considerably more than the population of an entire congressional district in this country, indeed closer to two district's worth. Arizona, meanwhile, was estimated in 2008 to have a population of 6,500,000 people, and is running at 48%, which means it's currently shorting itself by 130,000, which may not be a full congressional district's worth of people yet, but still ain't chicken-feed. Georgia, with a population estimate of 9,829,291 in July of 2009, is running 5% behind the national average, shorting itself 491,464 people, which happens to be inching right up to being as much as the entire population of Wyoming, which gets equally-weighted credit in Mr. Silver's analysis for being 3% ahead, as Georgia does for being as far behind as Wyoming's projected total.

In all Mr. Silver's analysis shows that the red states are actually running behind by a total of almost 1,500,000 people -- or roughly the entire populations of North and South Dakota put together.

And the thing is, folks, this is just a current "tally," it isn't even extrapolated to the mail-deadline.

Listen, I allowed a little too much personal envy of to cloud my discourse on the subject of what the very smart people over at FiveThirtyEight are accomplishing with their site -- especially when one considers that, first, it isn't a zero-sum game, and, second, even if it were, it can't be won by someone who takes a year off from playing -- and I got deservedly spanked for it. Mr. Silver and his team did a positively superlative job of covering the shocking absence of a McCain ground offensive during the 2008 campaign: it was a Peabody-worthy effort on their part, in the highest, grandest traditions of gumshoe journalism. Since then his coverage of the political strategies behind health insurance reform have been consistently incisive, often witty, and sometimes poignantly predictive of the drama's next chapter. He's a great writer, and a great economist. A pro in a world of hack amateurs, doing a peerless job of trying to keep the fractious liberal blogosphere on-task.

...But in an era in which every single statement of fact is carefully scrutinized by the (factless) other side for even the slightest hint of book-cooking, the fellow among us whose angle is nailing-down the numbers simply must be held to a higher standard than this most recent article.
He knows better.

Dave O'Gorman
("The Key Grip")
Gainesville, Florida.

No comments: