DiD Resources

I have done a lot of work on difference-in-differences (DiD), and have tried to make as many resources as possible available online. This page collects links to those resources.

DiD Course

I have given several one-day courses on recent DiD methods for various institutions and through Mixtape Sessions. All of my teaching material is available on Github.

This includes slides for three lectures – feel free to use in these in your teaching (with proper attritubution):

• Introduction (pdf latex): covers the canonical two-period DiD model and summarizes how recent papers have deviated from the canonical model

• Staggered treatment (pdf tex): covers (some of the) recent innovations in DiD with staggered treatment timing

• Violations of parallel trends (pdf tex): covers recent work on limitations of pre-trends testing and alternative approaches when parallel trends might be violated, with a focus on Roth (2022, AERI, “Pre-test with Caution…’’) and Rambachan and Roth (2023, RESTUD, “A More Credible Approach…”)

There are also two coding exercises that show how these methods can be used in practice:

• Staggered timing coding exercise: Instructions R solutions Stata solutions

• Violations of PT coding exercise: Instructions R solutions Stata solutions

Recordings for a virtual workshop I gave at the University of Michigan are available on Youtube: Part 1 Part 2

DiD Review Paper

For a more detailed review of the recent literature, see my Journal of Econometrics paper, “What’s trending in difference-in-differences? A synthesis of the recent econometrics literature”. You may be especially interested in Table 1, which provides a checklist for practitioners.

Also be sure to check out Table 2, which provides a list of packages for DiD methods, based on Asjad Naqvi’s excellent website.

Packages

I am the co-creator of several packages for DiD methods. Click on the links below and check out the package READMEs for instructions.

• HonestDiD (R Stata): robust inference and sensitivity analysis tools for DiD when parallel trends may be violated, based on Rambachan and Roth (2023)

• staggered (R Stata): Implements the efficient estimator for settings with (quasi-)random treatment timing proposed in Roth and Sant’Anna (2023, JPE:Micro). Also implements Callaway & Sant’Anna and Sun & Abraham estimators (without covariates).

• pretrends (R Stata Shiny): implements power calculations for pre-trends tests following Roth (2022, AERI)