Statistics, Machine Learning, and Understanding the 2016 Election

Statistics, Machine Learning, and Understanding the 2016 Election

-
Samuel Wang, Princeton University
Fine Hall 214

Although 2016 is a highly unusual political year, elections and public opinion follow predictable statistical properties. I will review how the Presidential, Senate, and House races can be tracked and forecast from freely available polling data. Missing data can be filled in using a Google-Wide Association Study (GoogleWAS). Finally, simple statistics can be used to identify inequities such as partisan gerrymandering, and provide a tool for possible judicial relief. These examples show how statistics and machine learning can deepen an understanding of the U.S. political scene, even under extreme circumstances.  Samuel S.-H. Wang, Ph.D., Professor, Neuroscience Institute and Department of Molecular Biology, Princeton University