Econometrics for Health Policy

GHP 525 Subject & Catalog Number

Description

This is a course in applied econometrics for doctoral and advanced master level students. The course has two primary objectives: (1) to develop skills in linking economic behavioural models and quantitative analysis, in a way that students can use in their own research; (2) to develop students' abilities to understand and evaluate critically other peoples' econometric studies.

The course focuses on developing the theoretical basis and practical application of the most common empirical models used in modern health policy research. In particular, it pays special attention to a class of models identifying causal effects in observational data, including randomized trials, instrumental variable estimation, interrupted time series, difference-in-difference, regression discontinuity, and propensity score methods. The course will also discuss approaches to dealing with missing data. We will also examine machine learning models for out of sample prediction, including random forest and gradient boosting methods.

Lectures will be complemented with computer exercises building on public domain data sets commonly used in health research, such as the Demographic and Health Surveys. Students will be expected to carry out a group project, using the methods taught in the course for a hypothesis and dataset of their own choosing. The statistical packages recommended for the exercises is STATA, R, or Python.

Course Note: Students are expected to be familiar with probability theory (density and distribution functions) as well as the concepts underlying basic ordinary least square (OLS) estimation.

Course Activities: Optional review and computer lab sessions will be held.
 
Course Prerequisites: BST210 or BST213; or equivalent course taken at Harvard Chan or HGSE with instructor permission