President Obama's re-election depended a lot on "The Cave." Meanwhile, a political historian accurately predicted Obama's victory almost two years ago, relying on questions he divined by studying volcanoes.
Time magazine was given an off-the-record, pre-election peek inside the Obama cave in Chicago where data crunchers and behavioral scientists analyzed reams of information — from who swooned at the sight of George Clooney to the most likely Obama supporters in Ohio.
As information trickles out, big data is being given a lot of credit for the targeting, fundraising and astonishing voter turnout that kept Obama in the White House for four more years.
Michael Scherer, the Time reporter allowed into the cave, said data analytics helped the campaign understand both the demographics they had tied down and the ones they needed to target.
Scherer said the operation called massive numbers of people nightly, working off a database of 29,000 people. He contrasted that to regular political polls that used a sample of 1,000 people. The larger sample and use of a database that allowed repeat calls gave the Obama campaign an ability to drill deep into the voter mindset heading to the election. It also gave them tips on how to get their voters to the polls.
Jim Messina, Obama's campaign manager, promised a metric-driven campaign where "we are going to measure every single thing in this campaign." The president's analytics team was five times larger than in 2008 and supervised by Rayid Ghani, whose experience included figuring out ways to maximize the efficiency of supermarket sales promotions.
The data crunching team worked in a windowless room, separated from other Obama staff. They gave code names to their various data crunching tests, which were described as the campaign's "nuclear codes" and not disclosed.
Using a unified megafile, rather than disparate databases, the team was able to generate loads of useful information, such as call lists for regional Obama campaign offices that listed phone numbers along with how they might be persuaded to support Obama.
"We could model people who were going to give through the mail. We could model volunteers," a senior team member told Scherer. "In the end, modeling became something way bigger for us in 2012 than in 2008 because it made our time more efficient."
Obama's cave men and women persuaded the campaign to wage a huge Facebook campaign, which in large part replaced door-to-door efforts. They also convinced senior advisers to let Obama field questions on the online, crowd-sourced news aggregator Reddit, even though older aides had never heard of it.
Database mining and social media outreach would be overkill for Allan Lictman who teaches at American University in Washington, D.C. He has correctly predicted the popular vote in the last eight presidential elections using a set of questions he devised by studying volcanoes and earthquakes.
"Everything we know about elections," Lichtman told NPR, "we've already stolen from geophysics. Tremors of political change, seismic movements of the voters, volcanic elections, political earthquakes."
The metaphors got him to thinking, so he borrowed earthquake prediction technology to analyze presidential elections between 1860 and 1980, looking for "markers of stability and markers of upheaval."
His research revealed that when six or more questions he developed went against the party in power, a challenger won. The questions range from whether the nation is in recession to whether there has been a milestone foreign policy success.
When he applied his 13 questions to Obama's re-election — in January 2010, he predicted, it turns out accurately, that only three questions went against the incumbent and he would be re-elected. The questions, or keys as Lichtman calls them, are contained in a book titled, "The Thirteen Keys to the Presidency."
If a party is in power, he says, it doesn't need a perfect record to be rewarded by voters. An incumbent can suffer up to five strikes before there is a "political earthquake."
NPR raised the question of whether Lichtman's predictive model conflicts with the way the news media — and campaigns — approach campaigning. "Primarily, elections are responsive to these deeper forces," Lichtman said. "Focusing on the campaign is like focusing on the froth of the wave, instead of the wave itself."