Peter Enns presents a district-level forecast of the 2026 midterm elections. The model would have correctly predicted the House majority in each of the last 14 elections. In addition to presenting the district-level forecast, the presentation will discuss the historical accuracy of the model, how polling data inform the predictions, and what the forecast can teach us about campaign dynamics and potential election anomalies. The model builds on Peter’s 2024 presidential election forecast, which correctly predicted the outcome in every state and forecasted Donald Trump’s popular vote within 1 percentage point.

Peter is co-Founder of Verasight; Professor at Cornell University; and Director of the Cornell Center for Social Sciences.

Registration coming soon