The Future of Predicting Elections


For the 2012 election, Nate Silver correctly called the result in all 50 states. How will his success affect 2016 and future elections?

In the days following the November presidential election, there was barely time to acknowledge President Obama’s successful electoral sweep before the idea that “it’s not too soon to start talking about 2016” became an immediate topic of discussion. Though it seems premature to start dissecting potential 2016 presidential candidates, it is by no means too soon to take a close look at a dominating factor in the 2012 race: election predictions.

Enter Nate Silver, a blogger and statistician who runs the FiveThirtyEight blog, which is licensed by the New York Times. Silver has a long history of successfully predicting the outcome of elections; in the 2008 presidential election, he called the winner of 49 of the 50 states, and this past November, he outdid himself by accurately predicting all 50.

Silver, though a Democrat, does this through a nonpartisan process that involves collecting state level polling data and averaging it, giving certain weight to recency, sample size, and the credibility of the poll itself. Once he calculates the weighted polling average, he takes voter adjustment and poll bias into account, and uses a linear regression analysis to see where the candidates stand.

Statistical models already play an important role in our day-to-day lives, and are used in just about everything from finance, to sports, medicine and weather forecasts. Though their use in election predictions is not exactly new — Yale’s Ray Fair created one of the first models designed to forecast elections — the prominence that Silver’s work has brought to polling analysis means that our awareness of this method and its implications is at an all-time high.

Admittedly, political pundits have been talking about the likelihood of election outcomes for as long as we can remember. However, the concrete nature of Silver’s numbers games — where else can you look at a calculation that not only accounts for all polls, but also considers their historical accuracy? — makes his straightforward, numbers-oriented approach undeniably appealing.

Yet this past week, in his “Ask Me Anything” session on Reddit, Silver acknowledged the probability of a “danger zone” created by predicting elections in this manner. On the day of the election, 27% of visitors to the New York Times site also viewed the FiveThirtyEight blog. That’s no insignificant number, given that the New York Times is the sixth-most visited news site in the US. In September 2012, there were 26 million visitors to the New York Times site. Assuming this number stayed relatively steady through the fall, this means that over 1 in 4 of these visitors stopped by FiveThirtyEight on Election Day to check out Nate’s predictions. That amounts to over 7 million people on November 6 alone.

The widespread awareness of Silver’s work can only be expected to grow significantly leading up to the next election. Yet the perception that “it has been foretold, therefore I won’t waste my time and vote,” is a problematic attitude to adopt, regardless of the candidate you’re backing. Suppose that on November 6, when President Obama supporters saw Silver’s prediction that their candidate had a 90.9% chance of winning, a number of them felt lulled into a false sense of security, and did not see the need to cast their individual vote.

On the other hand, it is possible that certain Governor Romney supporters saw Silver’s predictions that the Republican candidate had only a slim 9.1% chance of winning, and decided to not vote altogether, assuming that the election (from their perspective) was already lost.

Voter apathy is already a serious concern in American elections. Since 1960, the percentage of eligible voters who cast a ballot in a presidential election consistently falls somewhere between 49 percent and 63 percent. Conversely, in countries like Austria, Australia, and Belgium, the voter turnout rate reaches over 90 percent.

It’s also been said that the winner of American Idol gets more votes than our President. Though it’s important to remember that those American Idol voters can vote more than once, and do not have to be eighteen, the sentiment of this statement is still deeply alarming. How much more anti-motivation can we handle, and still keep participation in our elections afloat?

The transparent access to statistically-created election predictions doesn’t seem to have posed too much of a problem in the 2012 election, but as Silver’s work continues to gain widespread acceptance and offshoots of his “political calculus” inevitably spring up with other statisticians between now and 2016, it is probable that these election predictions will become a prominent part of upcoming elections.

The 2012 race, though an electoral landslide, was a close race in terms of the popular vote — only five million votes separated President Obama and Governor. That is two million fewer than the number of people who looked at Nate’s predictions on the day of the election.

As Silver’s popularity grows and the process of statistical election prediction becomes more heavily relied on, questions about the role of big data in our elections cannot be disregarded. In an age of rampant voter apathy, turning elections into transparent numbers games might harm much more than it will help.

About Adrian Rapazzini

Adrian Rapazzini is currently in her senior year at the University of California, Berkeley, where she studies English and Creative Writing. There are few things she loves more in life than a good chai tea latte and her cat, Kevin. Follow her on Twitter @adriana_rap.
Posted in: Culture, Election, Politics