DiDInt.jl

Welcome to the DiDInt.jl documentation!

DiDInt.jl is a package for implementing integrated difference-in-differences (DID-INT), a method developed by Karim & Webb (2025). DiDInt.jl allows for unbiased estimation of the average effect of treatment on the treated (ATT) in cases when the common causal covariates (CCC) assumption is violated. That is, it allows for unbiased estimation of the ATT when the effects of covariates on the outcome of interest vary by state, time, or both. Further, DiDInt.jl works for both staggered adoption and common adoption scenarios (assuming that once treated, always treated), and can report separate ATTs for each policy when treated states adopt different policies.

DiDInt.jl introduces two functions: one for estimation (didint()), and one for parallel trends and event study plots (didint_plot()). The syntax for these functions can be found here.

Some further details (a look under the hood) can be found on this page. Topics covered on this page include everything ranging from weighting options, aggregation methods, standard errors, computation of p-values via the randomization inference procedure described in MacKinnon and Webb (2020), and how DiDInt.jl processes dates and builds periods.

Examples can be found here!

Changes between versions of the package are listed in the changelog.