Huffpost Politics
THE BLOG

Featuring fresh takes and real-time analysis from HuffPost's signature lineup of contributors

Eric Zuesse Headshot

Revolutionary Study Shows Romney's Chances Are Overestimated In The Polls

Posted: Updated:
Print
AP
AP

A rigorous new scientific study from the Brookings Institution indicates that all nine of the leading predictions of the Presidential contest vastly overestimate Mitt Romney's chances of victory, even though all nine indicate that Barack Obama will probably win.

This study, titled "Forecasting Elections: Voter Intentions versus Expectations," shows that "voter expectations consistently yield more accurate forecasts than polls of voter intentions." The authors, David Rothschild of Columbia University, and Justin Wolfers of the University of Michigan, find: "In the 77 cases in which the intention ['For whom will you vote?'] and expectation question ['Whom do you expect to win?'] predict different candidates, the expectation question picks the winner 60 times, while the intention question only picked the winner 17 times."

All nine leading forecasts of the outcome of the U.S. Presidential election currently predict that Obama will probably win, but they're all based upon polls of voter intentions, which indicate much closer to an electoral tie than do polls of voter expectations. Polls of expectations have consistently indicated that huge majorities of voters expect Obama to win.

Perhaps the leading electoral forecast is Nate Silver's, and his current forecast at fivethirtyeight.com gives Romney only a 16.3% chance of winning, and it indicates, as a likeliest outcome, Obama winning around 305 Electoral College votes and Romney winning just around 233. However, Rothschild and Wolfers note, regarding the 2008 Presidential contest, that "forecasts based on prediction markets [which show expectations] yielded systematically more accurate forecasts of the likelihood of Obama winning each state than did forecasts based on aggregated intention polls compiled by Nate Silver for the website FiveThirtyEight.com." And that's just one example from this study's vast base of polls and predictions that were examined. But it's typical. Consequently, based upon all of the available evidence to date, Romney would have far less than the 16.3% chance of winning the White House that's indicated in the current forecast from Nate Silvers.

Rothschild and Wolfers theorize to explain why expectations-polls are enormously better predictors than intentions-polls. Their theory, boiled down, is: When you expect a particular candidate to win, that's a reflection of your entire circle of acquaintances, whereas when you intend to vote for a particular candidate, that's a reflection of only your own current intention - and, evidently, the broader sample of your acquaintances predicts better than does your current individual voting-intention.

The Rothschild-Wolfers paper is so solid in its methodology, and so overwhelming in its findings, that it will likely transform the way election-forecasting is done in the future. Or, as National Public Radio put it in a report on Saturday, "'For Whom Will You Vote?' May Be Wrong Question." Increasingly, the important question will be: "Who do you think will win?"

Investigative historian Eric Zuesse is the author, most recently, of DEMOCRATIC vs. REPUBLICAN ECONOMICS: NO CONTEST - Democrats Always Better, and of CHRIST'S VENTRILOQUISTS: The Event that Created Christianity.