Big Data Analytics Predict Obama Win

Prediction models point to Obama victory--but, as big data election forecasters learned in 2008, it's still too early to count Romney out.

Jeff Bertolucci, Contributor

September 27, 2012

6 Min Read

Social Studies: Obama vs. Romney

Social Studies: Obama vs. Romney

Social Studies: Obama vs. Romney (click image for larger view and for slideshow)

The latest polls show a tight 2012 presidential race, with incumbent Barack Obama maintaining a slight edge over challenger Mitt Romney. But according to two experienced election forecasters, President Obama is almost guaranteed a victory on November 6th.

Using two very different forecasting models, University of Illinois professor Sheldon Jacobson and American University professor Allan Lichtman have essentially reached the same conclusion: Unless a dramatic event occurs between now and Election Day to sway more voters toward the Republican candidate, a second term for President Obama is pretty much a lock.

Obviously, some Republicans may deride Jacobson and Lichtman as purveyors of partisan politics, but the professors insist their findings are based on sound analytics, not politics.

"We don't have any background in politics at all. We're completely neutral. We're not bipartisan, we're nonpartisan," said Dr. Jacobson, a computer science professor at the University of Illinois, in a phone interview with InformationWeek.

Jacobson and Lichtman will discuss their findings next month at the annual meeting of the Institute for Operations Research and the Management Sciences (INFORMS) in Phoenix, Arizona.

[ Semantic search sites offer straight answers on candidates' positions on the issues. See Big Data Sites Filter Election Noise For Facts. ]

Jacobson's Election Analytics Team at the University of Illinois takes the latest polling data and runs it through an advanced analytics model that incorporates Bayesian statistics to determine the probability of each candidate winning every state election. It then uses a dynamic programming algorithm to determine the likely number of Electoral College votes that each candidate will win on November 6. The model is used to predict Senate races too.

Jacobson's team updates the 2012 Election Analytics site as soon as new polling data becomes available. An earlier version of this model, the results of which are archived here, accurately predicted the outcome of the 2008 presidential election.

When interviewed on September 26, Jacobson said that if the election were held that day, Obama had a 100% chance of winning. Candidates' fortunes can change quickly, however, which was exactly what happened four years ago in the presidential race between Senator John McCain and then-Senator Obama.

On September 19, 2008, McCain had a 72% chance of winning the election, according to Jacobson's model. But by September 30, the probability of a McCain victory had plummeted to zero--and never recovered.

So what changed?

"McCain was leading quite soundly," said Jacobson. "If you think back to the end of September 2008, what happened was the financial crisis, and that was attributed to President Bush. Ultimately, all the people who were attached to President Bush, which included Senator McCain, ended up getting caught up in that, which led to his eventual demise in the election."

Allan Lichtman's forecast model is radically different. In a recent podcast on INFORMS' Operations Research website, the American University history professor called his prediction system "The Thirteen Keys to the White House."

According to his book, Predicting the Next President: The Keys to the White House, Lichtman first developed the system in 1981 in collaboration with Vladimir Keilis-Borok, a leading authority on the mathematics of prediction models.

"We developed it in 1981. It has since, in seven elections in a row, correctly forecast the popular vote outcome of the presidential election, usually months, sometimes even years, ahead of time. So it's got a 7-for-7, advanced prediction track record," Lichtman said in the podcast.

The model's premise is that elections are won or lost based on the performance of the party occupying the White House, not by what the challenger does. The "keys" are 13 factors that gauge the strength and performance of the party in power. They examine a baker's dozen of major issues, including foreign policy successes and failures, policy changes, social unrest, scandals, and third-party candidates.

"If the party holding power wins eight or more of the thirteen keys, they will be reelected. If they win fewer than eight, the challenger is expected to win," Lichtmann said.

Unlike Jacobson's forecast model, which is based on up-to-the-minute polling data, Lichtman's approach allows him to predict presidential elections far in advance. For instance, he first predicted the 2012 election in January 2010.

"Barack Obama has two keys more than the minimum necessary to win reelection, and that forecast has been stable for two-and-one-half years," said Lichtman.

The 13 keys, which are equally weighted, are clearly subjective. However, it's hard to argue with Lichtman's success at predicting presidential elections.

Here's a breakdown of Lichtman's 13 keys, and how President Obama fares on each.

Key 1. What were results of 2010 mid-term elections? Republicans win control of the House of Representative and gain Senate seats. Obama loss.

Key 2. Was there a serious challenge to the incumbent's party nomination? No. Obama win.

Key 3: Is the candidate the sitting president? Yes. Obama win.

Key 4: Is there a third party with a chance of getting 5% or more of the national vote? No. Obama win.

Key 5: Is the economy in recession during reelection year? Technically, no, although Lichtman acknowledges the recovery is limping along. Obama win.

Key 6: Strength of the long-term economy? Obama loss.

Key 7: Has the current administration undertaken major policy change? Yes, primarily healthcare reform. Obama win.

Key 8: Is there major social unrest in the U.S., such as what occurred in the 1960s? No. Obama win.

Key 9: Is there a scandal like Watergate or the impeachment of Bill Clinton? No. Obama win.

Key 10: Has the administration avoided a catastrophe abroad, such as losing a war? Despite the current global unrest, yes, it has avoided catastrophe. Obama win.

Key 11: Has the administration achieved a major foreign policy success? Yes, the "elimination" of Osama bin Laden. Obama win.

Key 12: Is the incumbent candidate charismatic? No. Obama was charismatic as a challenger, but he's lost his charm. "My most controversial call," said Lichtmann. Obama loss.

Key 13: Is the challenger charismatic? No, Romney is no charmer. "He's no Ronald Reagan," said Lichtmann. Obama win.

Final tally: Obama wins 10 of 13 keys.

In-memory analytics offers subsecond response times and hundreds of thousands of transactions per second. Now falling costs put it in reach of more enterprises. Also in the Analytics Speed Demon special issue of InformationWeek: Louisiana State University hopes to align business and IT more closely through a master's program focused on analytics. (Free registration required.)

About the Author(s)

Jeff Bertolucci


Jeff Bertolucci is a technology journalist in Los Angeles who writes mostly for Kiplinger's Personal Finance, The Saturday Evening Post, and InformationWeek.

Never Miss a Beat: Get a snapshot of the issues affecting the IT industry straight to your inbox.

You May Also Like

More Insights