Political statistics junkies will be buzzing today about Nate Silver’s “A ‘Radical Centrist’ View on Election Forecasting,” a characteristically-perceptive analysis at his Five Thirty Eight blog. It’s a response to a recent critique of Silver’s New York Times Magazine article, “Is Obama Toast? Handicapping the 2012 Election.” The critique, entitled “Why Data Wonks Are Wrong About Presidential Elections,” written by Ron Klain, a former Biden staffer, argues that statistical forecasts don’t help much with campaign strategy.
I haven’t yet read Klain’s post, so I can’t make a sound assessment about how accurately Silver describes his argument. But Silver’s post has merit even as a stand-alone essay on the nuances and concerns of political forecasting, one likely to be widely-discussed in poly sci classes this week.
I can certainly see good uses of forecasting in formulating strategy. Ad placement and timing, candidate travel, messaging and policy positions can all benefit from political forecasting. That’s not to say forecasting is the primary strategy tool for political campaigns, nor to deny that it’s been overvalued.
To a some extent, political forecasting is a sideshow of more immediate use to academics, journalists, gamblers and water-cooler chat than it is to political campaign workers. Accurately predicting which candidate wins gives political forecasters some cred as advisers and commentators. And, “If the election were held today” polls, more than forecasts, can tell a candidate that she/he is not doing well with union members, single women or suburbanites, for example. That can be helpful for tweaking messaging, ads and other campaign activities.
But Silver makes it clear that it would be folly for a campaign to marshal most of it’s resources in response to a forecast or poll and he is quite candid about the limitations and misuses of forecasting:
…Forecasters who are not conscientious about their methodology will wind up with models that make overconfident forecasts and that impute meaning to statistical noise…I would not paint all the forecasters with the same brush. Two political scientists who I know have a very sophisticated understanding of these problems are Larry Bartels at Princeton and Robert Erikson at Columbia. Others, like Hans Noel, will publish models, but provide very explicit disclaimers about their limitations. But there others who tweak as many knobs as they can, and there are bloggers and reporters who take all of the results at face value and don’t distinguish the responsible forecasts from the junk science. The problem is made worse when a game show is made out of forecasting and everyone competes to see who can get the most overfit model published in a peer-reviewed journal.
Silver weighs the relative influence of economic statistics, campaign strategy, candidate skills and ideology and unpredictable events, and discusses the difficulty of quantifying such factors. He notes that a lot of forecasting models flunked last year, and he cautions that the 2012 elections are likely to be especially problematic for forecasters. “We have already had an extremely wide array of outcomes in the various special and interim elections that have taken place around the country so far this year,” says Silver, “and we’ve had a very wild Republican primary, suggesting that voter preferences may be more malleable than normal.” He adds that even the best forecasting models don’t do such a great job of explaining the why of outcomes.
Alan I. Abramowitz, senior columnist for Larry J. Sabato’s Crystal Ball and member of the TDS Advisory Board, also has an insightful critique of Silver’s New York Times Magazine article, noting “several problems with Silver’s model,” among them,
…First, it isn’t really a forecasting model because the growth rate of the economy during the year of the election won’t be known until long after the election is over. In addition, the measure of the opposition candidate’s extremism is highly subjective…More importantly, Silver’s model may underestimate Barack Obama’s chances of winning a second term in the White House because it does not take into account the advantage enjoyed by first-term incumbents. And that advantage, as we have seen, is quite substantial.
But Silver does believe that political forecasting models can be made better and better, and can be increasingly useful in political campaigns. Anyone interested in data-driven political analysis should find Silver’s perspective a good read.