ABSTRACT
In this chapter, the authors introduce ideas behind the statistical models, also known as probability models, that were used by poll aggregators to improve election forecasts beyond the power of individual polls. Since the 2008 elections, other organizations have started their own election forecasting group that, like Nate Silver's, aggregates polling data and uses statistical models to make predictions. By understanding statistical models and how these forecasters use them, the people will start to understand how this happened. Forecasters also use models to describe variability at different levels. One of the most successful approaches used for this are hierarchical models, which can be explained in the context of Bayesian statistics. The authors briefly describe Bayesian statistics. Pollsters tend to make probabilistic statements about the results of the election. Forecasters like to make predictions well before the election. The predictions are adapted as new polls come out.
