Monday, October 31 & Friday, November 4, 2005
12:00 - 1:00 p.m.
ISPS, 77 Prospect Street, Room A002
(Lunch will be served in advance at 11:45 a.m.)
Click here to register.
These two lectures are designed to provide an introduction and tutorial on multilevel modeling. An attendee should come away with an intuition for what these models are, when they are most applicable, how they generalize to fit within the range of more commonly used models and how to run a multilevel model. Use will be made of new Bayesian technologies that have developed in recent years.
The first lecture on October 31 will provide a general discussion of the models. This will include their applicability and the advantages they offer. The second lecture on November 4 will provide a more practical approach toward understanding how to estimate these models (e.g., software, programming, etc.). A data set will also be made available for use in a hands-on exercise that may be conducted between or after the lectures.
Suggested background articles on multilevel models:
Marco R. Steenbergen and Bradford S. Jones, "Modeling Multilevel Data Structures," AJPS 46:1 (January 2002). Available at JSTOR
Bafumi et al., "Bayesian Multilevel Estimation with Poststratification: State-Level Estimates from National Polls," Political Analysis 12:4 (2004). Available at Oxford Journals
People interested in Bayesian estimation strategies specifically can read Simon Jackman, "Estimation and Inference via Bayesian Simulation," AJPS 44:2 (April 2000). Available at JSTOR
People interested in multilevel modeling as applied to ideal point estimation can read Bafumi et al., "Practical Issues in Implementing and Understanding Bayesian Ideal Point Estimation," Political Analysis 13 (Spring 2005). Available at Oxford Journals
For more information, contact Pam Greene, (203) 432-3052.
Joe Bafumi, Ph.D., Columbia University (2005), is a Postdoctoral Fellow at the Yale Center for the Study of American Politics. His primary research interests include American political behavior and public opinion. His articles have appeared in the British Journal of Political Science, PS: Political Science & Politics and Political Analysis. He is the co-recipient of the Warren Miller Prize for best article in the 2004 edition of Political Analysis. He has taught courses in American politics and methods at Queens College, Barnard College, and Columbia University. He will join the faculty at Dartmouth College in July 2006.
Sponsored by the
Institution for Social and Policy Studies