The Polling Observatory team is a polling prediction organization made up of professors and lecturers from various American and British universities. The latest significant addition to the analyst team was Christopher Wlezien of the University of Texas at Austin, who helped provide a detailed forecast for 2015.
The forecasting method used by the Polling Observatory team is based on polls from previous years, which allow you to see the dynamics of changes in public opinion. Thus, the basis for the prediction for 2015 was the survey data starting from 2010.
The first stage of the study is to analyze the electoral cycle and the willingness of people to vote for one of the three main parties. For the most accurate estimate, the organization used more than 3,000 polls since 1945, which provided a complete picture of public opinion.
The second step is the standardization of analytics and the so-called noise smoothing, which consists in averaging the data, allowing you to get a smoother graph without sharp drops.
The third step is to directly use the smoothed regression data obtained in the second step to estimate the likelihood of voters’ intentions and build a Bayesian poll averaging plot.
These three stages are key in predicting the 2015 elections by the Polling Observatory organization.