Stanford Seminar - How Can We Understand and Evaluate Election Forecasts, Given That N=15 or Less
Stanford University via YouTube
Free courses from frontend to fullstack and AI
AI Engineer - Learn how to integrate AI into software applications
Overview
Build a Learning Habit
Download Class Central's free printable study calendar
Download for Free
Explore the complexities of election forecasting in this Stanford seminar featuring Andrew Gelman from Columbia University. Delve into the challenges of understanding and evaluating election predictions with limited data points. Learn about preelection polls, differential nonresponse, external calibration, and political polarization. Examine the intricacies of interpreting forecasts, including the significance of polls above the 95% interval. Analyze the shortcomings of popular forecasting models like 538 and SimMap, and understand the concept of National Swing in electoral predictions.
Syllabus
Introduction
Preelection polls
Election forecasts
Polls
Differential nonresponse
External calibration
Political polarization
How to understand a forecast
Polls above the 95 interval
How we lost
Problems with 538
SimMap
National Swing
Taught by
Stanford Online