Looking Back on the 2024 Election #

Data Debrief: Predicting the 2024 Presidential Election. This Northwestern-University-hosted event was held on Tuesday, January 14, 2025, and moderated by Mollie Foust: Listen to the Recording.

The Virtual Tout® issued its Final Forecast with the polls opening on the East Coast at 6 AM Eastern Standard Time, November 5, 2024:

• Republican: Trump/Vance (280)
• Democratic: Harris/Walz (258)

The actual 2024 electoral vote totals:

• Republican: Trump/Vance (312)
• Democratic: Harris/Walz (226)

We correctly predicted a Republican victory. But we underestimated the strength of that victory. On election day itself, 1.8 million shares were traded on contracts for the two major parties in the market that we relied on for the 2024 presidential election (PredictIt).

Some might argue that the winner of the 2024 election has been the prediction markets themselves, which have again proven to be more effective predictors of election outcomes than polls and pundits.

Price Movements in Prediction Markets #

In-depth analysis of prediction market prices reveals the strength of the Republican victory. We provide documentation in the news log for January 7, 2025. In forecasts for future years, we expect to incorporate more extensive analysis of prediction market prices.

Across mainstream media, independent media, and social media, we also seek a better understanding of the influences on public opinion, prediction markets, and voting behavior. We will devote more time to Media Research between now and the elections of 2026 and 2028. Media research may reveal leading indicators (media measures) for predicting prediction market prices.

What have we learned from the election of 2024 and recent events? That we still have much to learn. See Learning.

The Virtual Tout® provides quantitative analysis of prediction markets with a focus on political and economic events. We post research results, forecasts, and commentary on this site.

For up-to-the-minute forecasts and analysis, see TinyTicker™.

If you have to forecast, forecast often. Edgar R. Fiedler

Acknowledgements #

The Virtual Tout® thanks Shree Bharadwaj, data scientist and angel investor, for his support of research and development on prediction markets and surveys.

Northwestern Analytics Partners, LLLP, Andrew D’Amico and Michael Purvis, provided invaluable feedback on this website. They also joined Tom Miller in developing the TinyTicker™.

Thanks to Candice Bradley, Northwestern data science instructor and erstwhile anthropologist and journalist, who provided valuable feedback about this website and offered opinions and suggestions along the way, some of which were incorporated into the narrative.

Northwestern University’s School of Professional Studies and Media Relations, Global Marketing and Communications, have supported this work over the years, hosting public events about political research and data science, working with the media, and issuing press releases about election forecasting methods.