Boston Chapter of the American Statistical Association
Short Course in Western Massachusetts
Basic Concepts of Statistical Inference for Causal Effects in Experiments and Observational Studies
Donald B. Rubin (Harvard University) and
Elizabeth A. Stuart (Mathematica Policy Research)
Date & Time Saturday October 22nd, 2005
9:00 AM – 9:30 AM Check-in
9:30 AM – 5:00 PM Course
Location Seelye Hall Room 201
Cost Registration is $45 for BCASA members, $65 for non-members, and $20 for students (copy of ID must be sent with your advance registration). This will include the cost of the course, morning coffee, lunch, and course materials.
Registration Limited to 50 participants. Mail a check (along with your name and e-mail address) for the course fee, payable to BCASA, addressed to BCASA, c/o Nicholas Horton, Smith College, Clark Science Center, Northampton, MA 01063-0001. Pre-registration is required. Registrations will be accepted until the course fills, but should arrive no later than October 15th. No refunds are available after October 15th though registrations are transferrable. Receipts will be available at the event. Inquiries can be sent to nhorton at email.smith.edu or 413-585-3688.
Directions See http://www.smith.edu/about_visit_directions.php for directions to the Smith College campus. Parking is available in the garage on West Street (see http://www.smith.edu/map/ for a campus map).
Abstract This course will present the Rubin Causal Model perspective on understanding and teaching statistical inference for causal effects through potential outcomes. There are three parts to the course. The first part establishes the primitives that form the foundation. The second part presents inference based solely on the assignment mechanism; this perspective generalizes Fisher's (1925) and Neyman's (1923) randomization-based methods, and emphasizes the central role of the propensity score (Rosenbaum and Rubin, 1983). The third part presents inference based on predictive models for the distribution of the missing potential outcomes, formally, Bayesian posterior predictive inference (Rubin, 1978). In practice, the predictive approach is ideal for creating statistical procedures, whereas the assignment-based approach of Fisher is ideal for traditional confirmatory inference, and the assignment-based approach of Neyman is ideal for evaluating procedures. For best practice, being facile with all three approaches is important. There is essentially no prerequisite knowledge for this course, as the material is based on an introductory course taught at Harvard University and designed for students with very little quantitative background. Examples are presented from a variety of fields, including medicine, education, and economics.
Instructors Donald Rubin is the John L. Loeb Professor of Statistics at Harvard University. His research interests include causal inference in experiments and observational studies, inference in sample surveys with nonresponse and in missing data problems, application of Bayesian and empirical Bayes techniques, and developing and applying statistical models to data in a variety of scientific disciplines. Elizabeth Stuart is a researcher at Mathematica Policy Research in Washington DC. She received her PhD from Harvard University. Her research interests include matching methods in causal inference, use of administrative records in census data, and use of historical patient data in clinical trials. Versions of this workshop have been presented at the University of Wisconsin Medical College, the University of Minnesota, Harvard University, the Joint Statistical Meetings, the Food and Drug Administration, and the Karolinska Institute in Stockholm, Sweden.
Background Reading While not required, the following papers may be helpful to provide additional background for the course:
Holland, P. (1986). "Statistics and Causal Inference", with discussion and rejoinder. Journal of the American Statistical Association, 81, 945-970.
Little, R.J., and Rubin, D.B. (2000). "Causal effects in clinical and epidemiological studies via potential outcomes: Concepts and analytical approaches." Annual Review of Public Health, 21, 121-145.
Reiter, J. P. (2000) "Using statistics to determine causal relationships." The American Mathematical Monthly, 107, pp. 24-32.
last modified July 18, 2005