“Why Do We Still Estimate Linear Models and Should We Move On (to highly non-parametric models)?” with Neal Beck and Jason Guo, NYU

MACMILLAN-CSAP QUANTITATIVE RESEARCH METHODS WORKSHOP

Abstract: Machine learning is great for simple atheoretical predictions but is it useful in standard garden variety social science (in place of linear regression like models). Here we discuss some tools in the context of Bayesian Additive Regression Trees. We apply the method to a study of civil war outbreaks; some results are promising, others problematic.

Nathaniel Beck is Professor of Politics at NYU where he has been for the last 15 or so years. Before that he was at UCSD. His interests are in methodology and its applications wherever.

Jason Q. Guo is a PhD candidate in the Department of Politics at New York University. He studies comparative politics with a special focus on historical development of political and economic institutions. His current projects on comparative politics include the development of state capacity and the persistence and decline of political elites in historical China. He is also interested in quantitative methodology, and is working on the application of machine learning methods in conflict prediction.

This workshop series is being sponsored by the ISPS Center for the Study of American Politics and The Whitney and Betty MacMillan Center for International and Area Studies at Yale with support from the Edward J. and Dorothy Clarke Kempf Fund.