Sam Wang, Associate Professor of Molecular Biology and the Princeton Neuroscience Institute, maintains an interesting range of research interests. Within his academic specialties, biophysics and neuroscience, he uses probability and statistics to analyze complex experimental data. His laboratory uses advanced optical methods to monitor and manipulate function in the living brain. He is well known for his many academic papers that use this approach and now for the popular book: Welcome to your Brain: Why You Lose your Car Keys but Never Forget How to Drive (Bloomsbury, 2008).
But Dr. Wang may be best known by the public for his blog, the Princeton Election Consortium. His meta-analysis of the electoral vote predicts who would win the election were it held today. He maintains that the method provides more objectivity and precision than looking at a few polls.
That doesn’t mean that his calculation on October 3 (Obama 349; McCain 189) meaningfully predicts what will happen on November 4, but the statistical work will come to reflect the actual result of the election more closely as Election Day approaches. The analysis rests on the principle that the median of multiple state polls closely predicts actual electoral outcome.
He emphasizes that in 2004, his calculation precisely captured the final outcome, although he had incorrectly assumed that undecided voters would tend to vote against the incumbent, a tendency noted in pre-elections polls. Leaving out that assumption made the model’s prediction precisely correct.
In many respects, Wang notes, his model has become more sophisticated. His group now tracks all 50 states and the District of Columbia rather than simply the battleground states. In 2004, Wang invented the concept of the Jerseyvote, which measures the power of individual votes in each state to sway an election. As it turns out, the power of voters in battleground states markedly outweigh those of us who vote in Jersey. According to Wang’s scale, as of today, a vote in Ohio is worth approximately 24 times a vote in New Jersey.
Wang used early October polling data to produce the following distribution of outcomes. Wang prefers not to predict the outcome of the election. Fivethirtyeight.com uses a similar methodology but also offers a prediction on the eventual election winner. According to Wang, here is the current distribution of electoral votes:
Useful links:
Posted by Lorene Lavora
Sam made the New York Times
You can also take part in a Princeton Election Forecast being conducted by University researchers.