Selection Bias and the Experimental Ideal

Overview

This module introduces the potential outcomes framework that we’ll use throughout the course as well as the concept of selection bias before explaining how average treatment effects can be estimated when treatment status is randomly assigned.

Readings

Mastering Metrics: Chapter 1

J-PAL: A Balancing Act

Optional Readings

Impact Evaluation in Practice, first edition: Chapter 3

Use the link to access the French, Portuguese, and Spanish versions.

Price Subsidies, Diagnostic Tests, and Targeting of Malaria Treatment: Evidence from a Randomized Controlled Trial

Lecture Slides

A handout version of the lecture slides is available here.

Empirical Exercise

The in-class activity as a do file or pdf

The empirical exercise as a do file or pdf

A web version of the empirical exercise is available here.

Video Lectures from 2021

Lecture 2.1 The Potential Outcomes Framework (26:44)

Review Questions

1. What is the fundamental problem of causal inference?
2. What are potential outcomes, and how do they create a missing data problem in program evaluation?
3. What is selection bias? When and why is it likely to bias estimates of program impacts?

Lecture 2.2 The Experimental Ideal (13:17)

Review Questions

1. How can random assignment eliminate selection bias?
2. What is the Law of Large Numbers, and why is it important in randomized experiments?
3. How can we estimate the causal impacts of a program when treatment assignments are randomized?

Lecture 2.3 A Short History of Randomized Experiments (27:35)

Review Questions

1. Who was Ronald Fisher, and how did he contribute to the development of randomized experiments?
2. When were randomized trials first used in medicine? When were they first used in the social sciences?
3. What were the first randomized evaluations used in the international development context?

Lecture 2.4 Analysis of Randomized Experiments (8:32)

Review Questions

1. When treatment is randomly assigned, how can we test the null hypothesis thta the average treatment effect is equal to zero?
2. How can linear regression be used to analyze data from randomized experiments?

Further Reading

Papers Mentioned in the Lectures

The Design of Experiments, Chapter 2

The Entry of Randomized Assignment into the Social Sciences by Julian Jamison

Do Conditional Cash Transfers Improve Child Health? Evidence from PROGRESA’s Control Randomized Experiment by Paul Gertler

Worms: Identifying Impacts on Education and Health in the Presence of Treatment Externalities by Edward Miguel and Michael Kremer

Randomistas in Development Economics

The Poverty Lab (A New Yorker profile of Esther Duflo, from 2010)

A Nobel Prize for the Randomistas (a Center for Global Development blog post by me)

More Recent Evidence on Health Insurance

Health Insurance and Mortality: Experimental Evidence from Taxpayer Outreach

On the “Credibility” Revolution in Empirical Microeconomics

The Credibility Revolution in Empirical Economics: How Better Research Design Is Taking the Con out of Econometrics by Joshua D. Angrist and Jörn-Steffen Pischke

Better LATE Than Nothing: Some Comments on Deaton (2009) and Heckman and Urzua (2009) by Guido Imbens