Treatment Effects Analysis

A 2-Day Seminar Taught by Stephen Vaisey, Ph.D. 

Read reviews of this course

This course offers an in-depth survey of a family of techniques known as treatment-effects estimators. Treatment-effects analysis is a quasi-experimental technique for estimating causal effects from observational data using the potential outcomes or counterfactual framework. These techniques — which include propensity-score matching, inverse probability weighting, and “doubly-robust” estimators — are now widely used in the social sciences, health sciences, and public policy.

The goal of treatment-effects analysis is to identify the causal effect of a treatment on an outcome, such as the effect of a college education on earnings, the effect of divorce on child outcomes, or the effect of a training program on employee productivity. A major advantage of treatment-effects techniques over standard regression methods is that they can produce different estimates of causal effects for subjects who are likely to receive the treatment and for those who are unlikely to receive it, an important distinction for policy work.

This seminar will take participants from simple exact matching to recent developments like coarsened exact matching and doubly-robust estimators. Participants will get extensive practical experience by working through case studies from economics, sociology, medicine, and public health.

Though the seminar will focus on hands-on understanding, we will also use causal graphs to look more deeply into the assumptions required to achieve unbiased estimates. Participant will learn to see how these techniques can be used in their own research.

We will cover a variety of topics including exact matching, propensity score matching and weighting, other forms of non-parametric matching and weighting, regression adjustment, and various forms of doubly-robust estimators. We will also consider tests for violations of assumptions and ways to test the sensitivity of results to violations of untestable assumptions. Although we will focus primarily on binary treatments, we will briefly explore how these techniques can be applied to multivalued treatments as well.

Who should attend?

This course is for any who want to learn to apply this family of techniques to observational data. Participants should have a basic foundation in linear and logistic regression.

Location, Format, and Materials

The seminar meets Thursday, June 11 and Friday, June 12 at Temple University Center City, 1515 Market Street, Philadelphia, PA 19103. 

The class will meet from 9:00 to 4:00 each day with a one-hour break for lunch at noon.

Participants will receive a bound manual containing lecture notes, examples of computer code and results, and other useful information. This book frees participants from having to take notes during the seminar.


The instructor will use Stata (and some R) for examples but the course material is not dependent on any particular software package. Those who have Stata 13 or higher installed on their laptops are welcome to follow along with the examples but this is NOT required or necessary.

Power outlets will be available at each seat.

Registration and lodging

The fee of $995.00 includes all seminar materials. 

Lodging Reservation Instructions

A block of guest rooms has been reserved at the Club Quarters Hotel, 1628 Chestnut Street, Philadelphia, PA at a special rate of $147 per night. This location is about a 5 minute walk to the seminar location. In order to make reservations, call 203-905-2100 during business hours and identify yourself by using group code STA610. The room block will expire when it is full or on May 11, 2015. 


1. The potential outcomes framework
      a. The experimental ideal
      b. Quasi-experiments and self-selection
      c. Directed acyclic graphs (DAGs)
      d. ATT: average treatment effect on the treated
      e. ATU: average treatment effect on the untreated
       f. ATE: average treatment effect
2. Exact matching
     a. Stratification
     b. Weighting
     c. Requirements for estimating ATT/ATU/ATE
     d. Other assumptions
3. Propensity score matching
     a. P-scores as a solution to sparseness
     b. Estimating the selection model
     c. Matching on propensity scores
     d. Calipers
     e. Testing overlap assumption
4. Propensity score weighting
     a. “Missing data” motivation
     b. Comparing weighting and matching
5. Non-parametric matching
     a. Nearest-neighbor (Mahalanobis distance) matching
     b. Coarsened exact matching
6. Regression adjustment
     a. Balancing vs. conditioning
     b. DAG motivation
7. Doubly-robust techniques
     a. DAG motivation for DR techniques
     b. Combining weighting and regression adjustment
8. Overview of advanced topics
     a. Sensitivity analyses
     b. Extensions to multivalued treatments 


“This is a very practical and easy to digest course for people who are interested in moving into the area of matching and treatment effects. The code and course materials alone are worth the price, the excellent instructor is a bonus.”
  Kathryn Nowotny, University of Colorado

“This course was very helpful in not only providing the information I needed to implement the methods, but also in explaining how to address challenges and how to write up and defend decisions.”
  Liz Lawrence, University of Colorado 

“This was my first time at a Statistical Horizons course, and I wanted to see how well these short courses could work for me as a way to address gaps. I found the course to be terrific, and stimulating, and plan to come to another.”
  June Tester, UCSF Benioff Children’s Hospital

“Steve took a tough subject and broke it into smaller parts of the whole, which added great dimension to the subject. Thanks!”
  Don Hunt, Georgia State University

“This is a fantastic class taught by a phenomenal instructor. I think any social scientist should take it.”
  Gino Cattani, Stern School of Business, NYU

“It is a very informative, useful class on treatment effect.”
  Pengxiang Li, University of Pennsylvania

“Excellent course, very easy to understand regardless of the domain you’re coming from.”
  Razvan Lungeanu, Penn State University

“The clearest and most practical presentation of these diverse methods. I certainly feel confident to implement these methods.”
  Lenny Lopez, Massachusetts General Hospital