Econometric Methods for Causal Inference

Epidemiologists and clinical researchers are increasingly seeking to estimate the causal effects of health-related policies, programs, and interventions. Economists have long had similar interests and have developed and refined methods to estimate causal relationships. Examples include difference-in-differences, instrumental variables, and regression discontinuity. This course introduces a set of econometric tools and research designs in the context of health-related questions. The course topics are especially useful for evaluating natural experiments — situations in which comparable groups of people are exposed or not exposed to conditions determined by “nature” (not by a researcher), as occurs with a government policy or a disease outbreak.


The objectives for this course are for participants to:

  • Identify opportunities for rigorous evaluation of natural experiments;
  • Enumerate the (oft unstated) assumptions involved in interpreting research that aims to draw causal inferences; and
  • Implement the research techniques covered in class.


Biostatistical Methods of Clinical Research I (BIOSTAT 200), Biostatistical Methods of Clinical Research II (BIOSTAT 208), and Epidemiologic Methods (EPI 203), or equivalent experience, are recommended. It is also expected that students will be comfortable using Stata software, which will be used throughout the course.


Course Director:

Justin White, PhD

Associate Professor, Institute for Health Policy Studies
email: [email protected]


Each week, new material is introduced via lecture and readings. Lectures will consist of discussions of course concepts, critiques of papers applying a research technique covered in class, and implementation of research techniques in Stata. These will take place on Fridays from 12:00 PM to 3:00 PM, Apr. 1 to Jun. 3. Students’ knowledge is assessed outside of class through weekly quizzes or problem sets.

  1. Weekly Quizzes
    Quizzes will be completed online, typically due seven days after each class. Late quizzes will not be graded except in extraordinary circumstances. The main purpose of the quizzes is to ensure that you have completed and understood the content of the assigned readings and in-class lecture for that week. You may not work on the quizzes with others.
  2. Problem Sets
    There are two problem sets, each due at the beginning of class on the due date. Late assignments will not be graded except in extraordinary circumstances. Assignments may be handwritten or typed but must be legible. You can work with others, but each student must turn in their own assignment and must acknowledge intellectual contributions of anyone who has assisted them.
  3. Original Empirical Paper (3-unit option only)
    The paper will be based on a topic of your choice, using data of your choice, and should apply a technique covered in this class. You are encouraged to discuss your paper with others, but, as always, you must acknowledge any intellectual contributions made by others. See the Box course folder for more detailed guidelines. Extra class sessions will be offered to help students in the 3-unit option to develop an empirical paper.

The schedule for the quarter shows dates and times for all activities.


Mastering 'Metrics: The Path From Cause to Effect by J. Angrist and J Pische. Princeton University Press. 2015.

Stata Statistical Software (Stata Corporation, College Station, TX) will be used; version 13 or higher is acceptable. A six-month student license for Stata/IC is the least expensive option that will be suitable to complete all course assignments, but Stata/SE is recommended for robust future use. The TICR Program has arranged for a sizeable discount for UCSF-affiliated personnel.
Books may be purchased either through the publisher or a variety of commercial venues (e.g.,

Students may find other textbooks useful to enhance their learning. Textbooks which discuss the material at a slightly less advanced level than our course include:

Causal Inference for Statistics, Social, and Biomedical Sciences: An Introduction by G. W. Imbens and D. B. Rubin. Cambridge University Press. 2015. (The most thorough and current statement of potential outcomes models for causal inference.)

Mostly Harmless Econometrics by J. Angrist and J-S Pischke. Princeton University Press. 2009. (The more mathematically advanced companion to Mastering ‘Metrics.)

Introductory Econometrics: A Modern Approach by J. M. Wooldridge. Cengage Learning. 6th edition. 2016. (A thorough, introductory treatment of a broad range of econometric applications.)

A Guide to Econometrics by P. Kennedy. Wiley-Blackwell. 6th edition. 2008. (Intuitive feel for econometric concepts, alongside more technical discussion.)

Microeconometrics Using Stata by A. C. Cameron and P. Trivedi. Stata Press. 2009. (A useful, though aging, primer on using Stata for econometrics; the book also has parallel content to their econometrics textbook.)

Health Econometrics Using Stata Partha Deb, Edward C. Norton, Willard G. Manning. Stata Press. 2017. (Overview of a variety of econometric modeling approaches in Stata, including many not covered in this course, e.g., GLM, count models, and a mass at zero.)

Other than textbooks, all course materials and handouts will be posted on the course's online syllabus.



Grading will be based on quizzes, problem sets and a paper:

Item Description % grade 2 units % grade 3 units
Weekly quizzes 6 quizzes 50% 20%
Problem sets 2 problem sets 50% 30%
Empirical paper 1 paper - 50%


UCSF Graduate Division Policy on Disabilities

To Enroll

ATCR and MAS students use the Student Portal

Students taking individual courses:

Course Fees
How to pay (please read before applying)
Only one application needs to be completed for all courses desired during the quarter.