Université de Montreal
Department of Political Science
Program: PhDSupervisor: Jean-François Godbout
Co-supervisor: Ruth Dassonneville
Title: Forecasting elections results : which models for parliamentary democracies?
Election forecasting is now a thriving discipline in the United States, where a large number of different models are being used to forecast the outcome of congressional elections or the fate of presidential candidates. In fact, during the last forty years, most forecasting models have been created for the American case. For their part, Westminster-style democracies have received very little attention. Although we can find a certain number of forecasting models for the United Kingdom and Australia (for example, Jackman and Marks 1994; Nadeau, Lewis-Beck and Bélanger 2009), there exist only one model to predict the incumbent party's vote share in Canadian federal elections (Bélanger and Godbout 2010) and, to our knowledge, none has been created for New Zealand. Due to the dominance of the Democratic and Republican parties in the United States, the vast majority of American models predict the two-party vote share, a measure that is almost irrelevant for countries where more than two parties compete for votes. Hence, the goal of my thesis is to develop a series of theoretically-driven models that can be used to predict the popular vote share (and seat share) of the main parties in the most important Westminster countries.
Joseph-Armand Bombardier scholarship of the Social Sciences and Humanities Research Council; MA-scholarship of the Fonds de recherche du Québec, Société et culture; PhD scholarship of the Fonds de recherche du Québec, Société et culture
Forecasting Dutch Elections: An Initial Model from the March 2017 Legislative Contests
Journal: Research and Politics
First Page: 1
Last Page: 7