Econometric Institute and Princeton University Press organize the intensive PhD-course:
"Analysis of Treatment Response for Decision Making"
Prof. Charles F. Manski (Nortwestern University)
June 2-4, 2004
Erasmus University Rotterdam
This intensive three-day PhD-course provides an in-depth overview of the analysis of treatment response with a special focus on providing the information required by decision/policy makers. An important practical objective of empirical studies of treatment response is to provide decision makers with information useful in choosing treatments. Often the decision maker is a planner who must choose treatments for a heterogeneous population. The planner might, for example, be a physician choosing medical treatments for a population of patients or a judge choosing sentences for convicted offenders. It is unrealistic to think that studies of treatment response can provide all the information that planners would like to have as they choose treatments. However, researchers can aim to improve treatment choice by addressing several questions: How should studies be designed in order to be most informative? How should studies report their findings so as to be most useful in decision making? How should planners utilize the information that studies provide? The complete course consists of 6 lectures of 2 hours, a morning and an afternoon session on each of the days. The major lectures will be given by professor Manski. In order to make these lectures accessible for a large, interested group of students, two introductory lectures and a lecture on topics related to Manski's work are given. Lecture 1 (Dr. Paap) starts with a general introduction on the concepts and ideas that are related to the analysis of treatment response. Lecture 2 (Prof. Manski) makes the broad case that analysis of treatment response should aim to inform treatment choice decisions. Lectures 3 (Dr. Donkers) and 4 (Prof. Manski) lay out the decision theoretic framework of Manski (2000, 2002). This assumes that the planner observes some covariates for each member of the population-to-be-treated; for example, a physician may observe a patient's demographic attributes, medical history, and the results of diagnostic tests. The observed covariates determine the set of treatment rules that are feasible for the planner to implement. Each member of the population has a response function, which maps treatments into a real-valued outcome of interest. One assumes that the planner wants to choose a treatment rule that maximizes the population mean outcome; in economic terms, the planner wants to maximize a utilitarian social welfare function. Under these assumptions, the optimal treatment maximizes the mean outcome conditional on the person's observed covariates. Hence studies of treatment response are useful to the degree that they enable the planner to learn how mean outcomes vary with treatments and covariates. Lecture 5 (Prof. Abbring and dr. Van der Klaauw) provides a detailed application on the estimation of the treatment response. Lecture 6 (Prof. Manski) examines the implications for treatment choice of pervasive identification problems in studies of treatment response. The largest part of this lecture is devoted to the selection problem as it arises in observational studies and in randomized experiments with partial compliance. This lecture ends with a discussion on the statistical inference required by planners wanting to use study findings for the entire study population, not for only a finite sample of subjects.
Bas van der Klaauw