About Us

Background About Stanford Predicts

Stanford Predicts is a Non-Partisan site. We are solely interested in predicting the likelihood of either candidate winning, for purely scientific purposes. Please go to our Methodology page if you want to learn more about how our statistical analysis works, or to Polling to learn about how we aggregate different poles to predict the outcomes inside swing states.

A simplified version of the US Presidential Election problem was given as a take-home exam in MS&E 120 Probabilistic Systems Analysis. An “academic-industrial-grade” system was built to track the 2000 presidential election. The “project” was carried in a small scale with a handful of students. In the 2004 election cycle, we are expanding the participation to involve more students so that they will:

  • Learn more about the election process in the US
  • Help track the battle for electoral votes – state-by-state, particularly in the hotly contested states
  • Learn what “swing states”, “dead heat” really mean
  • Critically learn the methodology behind various polling taking organizations
  • Learn how to transform available data into useful input to probabilistically track the progress in this election
  • Learn the model building process – how to break down a large/complex problem into manageable parts and then put the parts back together consistently
  • Learn about probabilistic modeling
  • Above all, be informed of the election process and the “arts” of projection

Students form small teams to manage various components of the project. The project TEAM will meet once a week to report progress, coordinate various inputs/outputs, disseminate results. All participants will have an overview of the project even though each team focuses on one particular component. The following tasks/teams will be formed:

  • Web site maintenance
  • State-by-state tracking
  • Project coordination and data aggregation
  • Poll methodology research
  • Public Relations