Predicting elections is getting harder. What did the midterms teach us?
Nov 5, 2026 · 12m
More Science podcasts · More News podcasts
Summary
Political scientist Jack Remenda discusses the accuracy of his Senate election forecast with host Regina Barber, noting that while his model correctly identified most competitive races, it underestimated the Democratic overperformance. The conversation explores the evolving science of polling, including the challenges of non-response bias and the blending of historical fundamentals with recent data.
Topics discussed
Sponsor: Open Door
Jack Remetta's final pre-election forecast
Sponsors: Whole Foods, Capella, Mint Mobile
Reviewing the accuracy of the Senate predictions
Challenges in modeling: fundamentals vs. polls
The impact of non-response bias on polling
Why different forecasting models vary
The rise and risks of prediction markets
The nationalization of American politics
Credits and closing sponsors
More from Short Wave
Listen ad-free on Castria