Improving election prediction internationally

• DOI: 10.1126/science.aal2887
• Publication Date: 2017-02-02T19:15:18Z
• Authors: Ryan Kennedy; Stefan Wojcik; David Lazer
• Keywords: Predictive power; Federal election; Ranging

Using global data for election predictions Assumptions underlying result have been questioned recently. Kennedy et al. assessed more than 650 executive office elections in over 85 countries and performed two live forecasting experiments. They analyzed a variety of potential predictors theorized to be importance, ranging from economic performance polling data. Elections were about 80 90% predictable, despite uncertainties with available Polling very important successful prediction, although it was necessary correct systematic biases. Unexpectedly, indicators only weakly predictive. As sources improve grow, predictive power is expected increase. Science , this issue p. 515