Forecasting the 2012 and 2014 Elections Using Bayesian Prediction and Optimization

This article presents a data-driven Bayesian model used to predictthe state-by-state winners in the Senate and presidential elections in 2012 and 2014.The Bayesian model takes into account the proportions of polled subjects who favor eachcandidate and the proportion who are undecided, and produces a posterior probabilitythat each candidate will win each state. From this, a dynamic programming algorithm isused to compute the probability mass functions for the number of electoral votes thateach presidential candidate receives and the number of Senate seats that each partyreceives. On the final day before the 2012 election, the model gave a probability of(essentially) one that President Obama would be reelected, and that the Democrats wouldretain control of the U.S. Senate. In 2014, the model gave a final probability of .99that the Republicans would take control of the Senate