Nate Silver is a brilliant statistician. But he's not clairvoyant -- nor is he crazy -- and you'd have to be one or the other to attempt to forecast the 2012 presidential election with confidence.
That's why in Sunday's New York Times Magazine article entitled "Is Obama Toast? Handicapping the 2012 Election," Silver makes bold predictions but backs away quickly.
In polling, a year is an eternity. In November 2007, the Real Clear Politics average had Rudy Giuliani at 30.1 percent; John McCain, 16.4 percent; Fred Thompson, 16 percent; Romney, 11.9 percent; and Huckabee, 8.1 percent. It wasn't until January 11, 2008, that John McCain took the lead in the Republican primary for the first time -- with just 25 percent of Republican primary voters backing him.
A little less than a year before the 2008 election, Gallup found Hillary Clinton led now-President Barack Obama by 27 points. Clinton edged out all Republican contenders while Obama just barely tied with leading Republican candidates Giuliani and McCain.
Silver concedes he uses just 17 elections (the post-1944 presidential cycles), discounts the reliability of historical precedent, admits the inconsistency of economic forecasts, and exposes prevailing methods for predicting electoral outcome through tables and rundowns as no better than hit-or-miss.
In other words Silver outlines just how tenuous the variables from which he's attempting to draw his bombastic conclusion are. The more of these admittedly untrustworthy variables he strings together, the less significant his result.
Yet Silver perseveres to produce the media-genic "17 percent chance of reelection" statistic that became the cover of the New York Times Magazine. Only midway through the essay is it clarified that Silver's prediction applies to Obama's odds when pitted against Romney in a stagnant economy.
Silver tries to quantify the ideological stances that candidates hold -- an impossible task in November 2011, particularly in such a inchoate Republican primary. He applies historian Marty Cohen's methodology to pinpoint candidates' ideology using "Congressional voting records and surveys of presidential historians." With the Republican field largely un-vetted -- and currently dominated by a contender with no political record -- this methodology means bubkes.
The Republican candidates haven't been subjected to the intense scrutiny the ultimate nominee will face. With six candidates still nominally in the race -- Bachmann, Cain, Huntsman, Paul, Perry, and Romney -- media have not focused nearly as much as they will on any one contender. News on these six, even presumed nominee Romney, remains topical and reactive. Reporters respond to issues as they arise -- either in the form of incidents like Perry's erratic New Hampshire performance or Cain's ongoing sexual harassment blow-up.
Republican candidates' statements and actions thus far have been about appealing to the far-right base -- not accurately representing their past or future, general election ideology. Coverage of these candidates is usually about these attempts to appeal to the Republican electorate -- and Republican infighting. Wider public discussion and debate of their records and potential presidential decision-making style relevant to the general electorate won't happen until the primary ends.
There are significant unquantifiable factors in play in 2012. The eventual Republican nominee will have to run on Bush's and Boehner's records in the general election as well as his (or her) own. Following Obama's accomplishments and promises kept in Afghanistan, Iraq, and Libya national security is off the table. The last time Republicans lost national security as a campaign issue, a little-known Arkansas Democrat defeated a popular sitting Republican president.
The Obama campaign, Democratic National Committee, and progressive independent expenditures haven't even entered the fight yet. The White House isn't in campaign mode; it's governing. The ads and emails released thus far represent nothing more than a few shots across the bow.
Yet Obama remains ahead of or tied with Republican challengers facing less scrutiny and carrying minimal baggage -- two conditions that will change as soon as a nominee is named. In swing states (and should-be red states) like Colorado, Florida, Iowa, Michigan, New Hampshire, New Mexico, Nevada, North Carolina, Ohio, Pennsylvania, Virginia, and Wisconsin, Obama holds his own. Analysis of the electoral college map still yields routes to victory.
There's no way to reach a definite conclusion about what will happen a year from today using polling within the margin of error, economic forecasts, shaky quantification of candidate ideology, and presidential historians' analyses. It's not just likely that the variables on which Silver predicated his analysis will change -- it's certain.
Silver attempts to compensate for the certainty of change with hedging and concessions as to the unreliability of variables. What he doesn't do is rebut the common-sense conclusion -- backed by history and numbers -- that producing numbers in November 2011 is like trying predict the full season for both leagues of national baseball before the draft. There are simply too many variables.
To use a second sports analogy, Silver's attempt at qualifying the hard numbers and predictions he advances is like a soccer player protesting his innocence after fouling another player in public view. The best of intentions and excuses don't render a foul harmless. Disavowing responsibility for the consequences of his shaky statistical hyperbole doesn't make Silver any less culpable for its effect on the national political dialogue. There's also no excuse for perpetuating a view of polling and data analysis as either magic or exact science. It's neither.
Take for another example the Monday Washington Post article claiming that Mississippi's "personhood" amendment, which would declare zygotes, embryos, and fetuses legal "persons," was a sure bet. The article's ostensible evidence -- a Public Policy Polling poll showing 46 to 45 percent support in favor of the amendment -- didn't constitute proof or even strong support for the claim. The same article quoted Public Policy Polling Director Tom Jensen as saying that "Things can definitely go either way." In fact, the amendment failed -- with more than 55 percent of voters opposing its passage.
Silver's analysis is part of a larger problem: pundits and writers pushed to crank out catchy content produce analysis and prediction for mass consumption with little regard for whether it's right or wrong -- or even meaningful.