Skip to content

drdid

Fit a doubly robust difference-in-differences model to estimate the average treatment effect on the treated (ATT).

When to use

How do I estimate an average treatment effect on the treated under panel data, adjusting for covariates using outcome regression, propensity score weighting, or both (doubly robust)?

Syntax

drdid y [covariates], treat(<var>) post(<var>) [method(or|ipw|aipw) robust bootstrap(<n>) seed(<n>)]

Examples

panel firm_id year
drdid wage exper tenure, treat(treated) post(post)
drdid wage exper tenure, treat(treated) post(post) method(ipw)
drdid wage exper, treat(treated) post(post) robust bootstrap(100) seed(42)
estat drdid`  (post-estimation diagnostics: method, cell counts, propensity score summary)

See also