← All research spotlights

Andrew Gelman

Photo: Schutz, CC BY-SA 4.0, Wikimedia Commons

Research spotlight · Unit 6

Andrew Gelman

Columbia University

Research field
Bayesian statistics, survey methodology, political science

Andrew Gelman is Higgins Professor of Statistics and Professor of Political Science at Columbia University and works on Bayesian data analysis, multilevel models, and the methodology of surveys and elections. In a 2015 paper with Wei Wang, David Rothschild, and Sharad Goel he showed that a highly unrepresentative opinion poll, a daily question answered by hundreds of thousands of Xbox players before the 2012 presidential election, forecast the result to within a fraction of a percentage point once the responses were modelled by demographic cell and weighted back to the electorate. The unit presents this as the extreme case of what the weighting behind every published margin of error does.

In the masterclass

Section 4, after the margin of error and what a published margin assumes, shows how his team turned a poll of video game players into an accurate forecast.

The research is presented in Unit 6: Estimation and Confidence Intervals, in a research spotlight box that explains the result at the level of the unit and links it to what you have just learned.

The paper

Wang, W., Rothschild, D., Goel, S., and Gelman, A. (2015). Forecasting elections with non-representative polls. International Journal of Forecasting, 31(3), 980‑991. https://doi.org/10.1016/j.ijforecast.2014.06.001

The scientists named here are not contributors to the Statistics Masterclass, are not affiliated with it, and have not endorsed it. The program presents and explains their published research, with references to the original papers.