Proves Same Accuracy of Predictions for Targeted and General Population Samples
Preceding the January 31st Florida Republican Primary, Infosurv launched an iCE prediction market asking a general population of Americans the key question whose answer played a pivotal role in the 2012 election season: “Which GOP Presidential Candidate will win the historically crucial Florida primary?”
Recognized as one of the research industry’s most innovative new technologies, iCE (the Infosurv Concept Exchange) is a virtual online trading market, proven more accurate at predicting future outcomes than conventional surveys, polls, or individual experts. Infosurv adapted this virtual market design to predict the success of candidates in the 2012 political run for the White House – a novel ambition that was sure to reveal fresh insight into the applications for prediction markets in the political arena.
So what were the results?
iCE traders predicted GOP front runner Mitt Romney as the primary winner, followed by former House Speaker Newt Gingrich, then Rick Santorum and finally, Ron Paul. These predictions were proven extremely accurate. Not only did we predict the correct rank order, but our probabilities closely matched the actual margins in the election.
This parallel in distributions shows that a representative general population sample can be just as effective as a targeted sample in determining election results.
iCE predictions were also validated for their accuracy with the Iowa Caucuses and South Carolina Primary. Despite the volatile GOP environment leading up to the caucuses, the iCE market rankings aligned entirely to political polls of the same time period. Even more intriguing, the Iowa Electronic Markets, an online futures market operated by University of Iowa Henry B. Tippie College of Business faculty, reported the identical candidate rankings as did iCE. Authenticating the iCE methodology even further, our iCE market results for the South Carolina Primary aligned to political polls of that same time period as well. It was only due to late-breaking extraneous events (Perry’s withdrawal, announcement that Santorum actually won the Iowa caucus, and Gingrich’s defense of his ex-wife’s allegations during the final debate) that our iCE market predictions did not match the actual election.
Delving further into the voter psyche, iCE market data revealed which candidates were the most polarizing among potential voters, as well as which candidates may soon lose influence as a result. Our traders also provided qualitative insight into what each candidate could do to appeal more to voters. The feedback gave a perspective on each candidate that was both candid and constructive to those hoping to win the GOP nomination.
Contact us at [email protected] to learn more about the insights our political markets revealed and our plans for future enhancements for iCE.
