Speaker: Lenny Bronner, The Washington Post

Election Night Forecasting at The Washington Post

This talk introduces the methodology used by The Washington Post to forecast election results. Raw vote counts can be misleading due to factors such as early vote ballot processing, precinct reporting rates, and geographic voting patterns, so our goal is to offer readers a more accurate and nuanced understanding of the race as ballots are being counted. Our approach incorporates county and precinct-level election results, historical voting data, and demographic information to create a more accurate estimate of outcomes as votes are tallied. Valid uncertainty quantification for this high-stakes task is crucial. Using statistical techniques developed at Stanford University, such as the bootstrap and conformal inference, we generate prediction intervals that capture uncertainty from partial returns and demographic influences on voting behavior. Through this approach, we aim to provide a transparent, statistically sound analysis that moves beyond simple vote tallies to offer a deeper and more reliable perspective on the evolving electoral landscape.

Lenny Bronner is a lead data scientist with The Washington Post's Data & AI team, specializing in election data analysis and modeling. His work includes election night forecasting, poll aggregation, and voter registration analysis. Lenny holds a bachelor’s degree in Mathematical and Computational Science and a master’s degree in Statistics, both from Stanford University. Originally from Vienna, Austria, he now resides in New York City.

About the Seminar Series: Scientific Discovery in the Age of Data & AI

The last few years have seen a substantial increase in the reported success of machine learning (ML), and generative artificial intelligence (AI). These impact practices in the delivery of services from financial institutions to entertainment and medicine. But scientific research also increasingly relies on data analysis methods that can be described as machine learning and artificial intelligence. This seminar series aims to investigate if and how the paradigm for scientific research has changed or should change to incorporate these new tools and the possibilities they open.

A diverse group of scholars engaged in scientific research, method development, and historical and epistemological investigations will give a 50-minute presentation, followed by discussion.

The event is open to the Stanford Community. Stanford students and postdocs have the opportunity to engage more directly with speakers and topics by enrolling in the Canvas course here.

Speakers in the series include Ellen Zhong (Princeton), Tom Griffiths (Princeton), Thomas Icard (Stanford), and Jure Leskovec (Stanford) among others.