A new approach to determining which candidate will win the most electoral votes in the U.S. Presidential race factors in lessons learned from the 2004 election and uses sophisticated math modeling. The research will be presented at the annual meeting of the Institute for Operations Research and the Management Sciences (INFORMS®).
As of September 16, the margin in electoral votes could be as high as 282.8 votes for Senator John McCain against 255.2 for Senator Barack Obama, depending on the forecasting scenario.
Operations researcher Sheldon H. Jacobson, a professor at the University of Illinois at Urbana Champaign, along with a group of students and collaborators at Southern Illinois University-Edwardsville, created http://election08.cs.uiuc.edu/, a math model that dynamically forecasts the outcome of the election.
Prof. Jacobson and colleagues will present their findings at the INFORMS annual meeting, which takes place in Washington, D.C. at the Marriott Waldman Park Hotel and Omni Shoreham Hotel from October 13 – 15, less than three weeks before the election. Over 4,000 analysts and experts in analytics are expected to attend. Information is at http://meetings.informs.org/DC08/.
Jacobson's model applies a mathematical model to state polling data, using a dynamic programming algorithm to forecast electoral results.
"The results from the 2000 and 2004 presidential election suggested that it can be difficult to predict the winner of the presidential election based on popular vote," says Jacobson. "In fact, it is possible that the popular vote and the Electoral College vote can lead to significantly different results."
Jacobson's model employs Bayesian estimators (which help scientists make decisions when conditions are uncertain) to determine the probability that a candidate will win each state. He obtains state polling data from Rasmussen Reports, the Quinnipiac University Poll, and SurveyUSA. State-by-state probabilities are then used in a dynamic programming algorithm to determine a probability distribution for the number of Electoral College votes that each candidate will win in the 2008 presidential election.
Professor Jacobson believes that this model provides a more realistic method of predicting the results. In 2004, when most other polls showed Kerry with a clear edge, his model consistently showed a Bush victory.
"We take into account 'safe' states— states that each candidate is basically guaranteed to win," says Jacobson. "In 2004, once you took into account Bush's 'safe' states, he had a much narrower gap to close to get to 270 electoral votes than Kerry."
In the model, a safe state is one in which the candidate has an 85% chance or greater of winning.
Jacobson's model also factors in undecided voters. It accounts for five different voting scenarios involving undecided voters, each considered individually. A "Neutral" scenario provides an unbiased handling of undecided voters. "Strong Republican" and "Strong Democratic" scenarios provide two extreme envelopes around which results can be judged and evaluated, while "Mild Republican" and "Mild Democratic" provide more realistic possibilities if late-breaking information surfaces that shift voter preferences.
"Undecided voters can have a significant role on the outcome of the election. In fact, they are likely to be the ultimate deciders of who will win this election," says Jacobson.
Below is information about Prof. Jacobson's workshop at the INFORMS annual meeting:
A Prediction Model for the United States Presidential Election with Sheldon Jacobson, University of Illinois et. al. Sunday, October 12, 8-9:30 AM
Additional election-related workshops at the INFORMS annual meeting include the following:
- Political Engineering and Predicting the Outcome of the U.S. Presidential Election with James Cochran, Louisiana Tech University,
Wednesday October 15, 1:30 - 3:00 PM
- It's All Markov's Fault: Secrets of Targeting with consultant Doug Samuelson, examining candidate decisions such as where to appear, what issues to emphasize, how and where to get out the vote, what kinds of ads to run, and how to raise funds. Wednesday October 15, 1:30 - 3:00 PM
- Election Projections: A Retrospective with Jack Moshman, Moshman Associates, Wednesday October 15, 1:30 - 3:00 PM
- The Keys to the White House: 2008 Prediction with Allan Lichtman, American University. Sunday, October 12, 8-9:30 AM
- Panel Discussion: Should the Electoral College Stay, Be Reformed, or Be Abolished? with Arnold Barnett, MIT; Edward H. Kaplan, Yale; et al. Tuesday Oct 14, 8:00 - 9:30 AM
Doing Good with Good O.R. (operations research), a series of plenaries and panels about healthcare, energy/environment, and air transportation, will take place at the INFORMS annual meeting on Tuesday, October 14 and Wednesday, October 15.
About INFORMS
The Institute for Operations Research and the Management Sciences (INFORMS®) is an international scientific society with 10,000 members, including Nobel Prize laureates, dedicated to applying scientific methods to help improve decision-making, management, and operations. Members of INFORMS work in business, government, and academia. They are represented in fields as diverse as airlines, health care, law enforcement, the military, financial engineering, and telecommunications. The INFORMS website is www.informs.org. More information about operations research is at www.scienceofbetter.org.