On June 16, 2025, Rajiv Sethi, professor of economics, published new research titled “Political Prediction and the Wisdom of Crowds,” which investigates how well statistical models and prediction markets forecasted outcomes of the November 2024 U.S. elections.
Instead of using traditional accuracy measures, the researchers tested each method by imagining a trader who believes its forecasts and bets money based on this belief. If the model was good, the trader should profit. The statistical models failed to beat the prediction market for high-profile races like the presidency, where many informed people actively bet, but they outperformed the market in smaller, less visible races with fewer participants.
The paper is co-authored by Barnard alumna Julie Seager ’21, who is currently pursuing a PhD at MIT, alongside four current Barnard students—Anna Hammell ’26, Tianshuo Liu ’27, Sachi Patel ’27, and Ramya Subramanian ’26—who are Bridgewater New Pathways scholars. Their contributions were supported by the New Pathways grant, which funded their research collaboration with Sethi. The study will be presented at the 2025 ACM Collective Intelligence Conference and published in the forthcoming conference proceedings.