Forecasting election outcomes has become a popular sport among scholars, pundits, and the media alike. Statistical models that provide timely and precise forecasts are of great interest for campaign organizers and financial contributors who want to target their resources efficiently. In this course, we will study different methods of election forecasting and apply them in the context of upcoming elections.
In the first part of the course, we will critically discuss existing approaches of election forecasting, including (1) fundamentals models that exploit indicators of economic performance and political mood, (2) poll-based models that infer the result from public opinion surveys, (3) betting markets, and (4) approaches that draw on new data sources, such as social media data. In the second part of the course, we will turn to a practical implementation of these methods for upcoming elections, including (but not limited to) the 2017 German federal election. To that end, we will develop own statistical models, collect and prepare data and finally implement the forecasting using statistical software.
This course is for 2nd year MIA and MPP students only.