Causal Inference from Observational Studies: Difference in Differences (DiD)

Course details

statistics seminar | level: beginner | register now
for questions related to this event, contact ugent@flames-statistics.com
affiliation: Ghent University

Abstract

Difference in differences (DiD) is a quasi-experimental design that makes use of panel or longitudinal data from treatment and control groups to derive causal estimands. DiD is typically used to estimate the effect of a specific intervention or treatment (such as a passage of law, enactment of policy, or school program implementation) by comparing the changes in outcomes over time (at least between two periods) between those in the intervention group and those in the control group.
The DiD approach computes two types of differences: The first difference is the difference in the average of the outcome variable of interest between the two time periods for each of the groups whereas the second difference is the difference between the differences calculated for the two groups in the previous step. This second difference measures how the change in outcome differs between the two groups, which is interpreted as the causal effect estimate.
In this seminar, we motivate the need for the DiD approach, discuss the important assumptions and demonstrate how it works in R using data on a study of the effect of an incinerator on the value of homes closer to its site. We conclude with some issues with the DiD method and some extensions such as triple DiD.

Prerequisites

Background readings

Scott Cunningham . Causal Inference: The Mixtape. 2021

Myoung-jae Lee . Matching, Regression Discontinuity, Difference in Differences, and Beyond. 2016

Fee

Free

Venue

Online

Instructor

dr. Emmanuel Abatih