Predicting the 2004 Presidential Election

Yiqing Li successfully defended his M.Sc. thesis entitled "Predicting the 2004 Presidential Election" where he developed a method to help predict the upcoming US presidential election based on poll results from individual states.

Over the next few months, Americans will be repeatedly surveyed about their voting intentions in the upcoming Presidential elections. Although nationwide poll provide some indication of the mood of the American public, the polls ignore the role of the electoral college system in determining the election of the president.

In some cases, poll results are reported on a state-wide basis. In this thesis, a model is built that uses the state wide results and the margin of error of each state's results to make a probabilistic assessment of victory. Estimates can be readily updated whenever new surveys become available. In addition, a sensitivity analysis is performed to gauge the effect of shifts in public opinion for the candidates.

This type of interdisciplinary work is a hallmark of our program in Applied Statistics at Simon Fraser University. For more information, please contact Yiqing Li (ylia (at) stat.sfu.ca) or his supervisor Tim Swartz (tim (at) stat.sfu.ca).