Predicting the Next President: A Demographic Model

Abstract

Political scientists have created numerous forecasting models for elections worldwide. In the United States, there are several models that predict the outcome of the presidential election (see e.g., Abramowitz, 2016; Lewis-Beck & Tien, 2016; Lockerbie, 2016). Beyond models published in the scholarly literature, a host of models published through mainstream media sources have cropped up in the last decade as the Internet has become more popular and more accessible (see e.g., FiveThirtyEight, The Huffington Post, The New York Times). Despite the growing number of forecasting models, no models currently attempt to predict the results of the presidential election by predicting the turnout and vote of demographic groups. This project includes a background on presidential election forecasting with a review of the literature and our two-step model to predict the turnout and vote of four ethno-racial demographic groups. I completed a preliminary model and issued a prediction prior to the 2016 election. The project concludes with an analysis of the accuracy of this model and a comparison of the accuracy of other published models to determine whether my demographic model is useful to the future of election forecasting.