400001 SE SE Methods for Doctoral Candidates (2012S)

The course provides instruction in election forecasting. The word, forecasting, is sometimes used loosely. Here it will have a very specific meaning: the prediction of an event before it happens. Most social science, political science included, does not do forecasting. Instead, they engage in description, e.g., what is happening or what has happened, or explanation, e.g., what are the theoretical reasons something has occurred. Explanation, in particular, wears the scientific mantle, and the tools for its investigation can be very sophisticated, i.e., theories, hypotheses, concepts, measures, presented systematically in an equation with a dependent variable as a function of independent variables. That model, as it is called, may then be estimated by a statistical technique, most probably some form of regression analysis.

Election forecasting has taken place at all system-levels of the polity, from the local to the constituency to the region to the nation. Most of it, however, has focused on national elections, mainly presidential or parliamentary. That will be the emphasis here. Although election forecasting is a relatively new field, by now a number of studies have cumulated, at least for certain countries. Denmark, France, The United States and the United Kingdom are often the subject of such investigation. Note that these political systems have important differences, such as the executive arrangements or the number of parties. Thus, we observe that forecasting of elections need not confine itself to simpler systems, say with two-parties and a single executive. As well, Canada, Germany, Hungary, Italy, Portugal, and Spain have to their names reputable election forecasting investigations. And, the list of nations so studied is growing. For example, as a result of efforts by my students from earlier classes, we have now added to the world forecasting models from Belgium, Turkey, Columbia, Lithuania, Norway, to mention some. Undoubtedly, during this class, other forecasting models will be identified for other nations represented by students in the course.