did-analysis

To perform a difference-in-differences analysis, provide panel data containing entity_id, year, treatment, and outcome to trigger a workflow guiding assumption checks, model specification, and robustness tests. This produces credible causal inference estimates for policy changes or natural experiments.

To check the parallel trends assumption, the skill guides you through pre-treatment event-study diagnostics and assumption validation. This verifies pre-treatment outcome trajectories are consistent before estimating the actual treatment effect.

Yes, you can use Callaway-Sant'Anna, Sun-Abraham, and staggered adoption estimators alongside standard 2x2 DID and TWFE designs. The skill specifies appropriate models and calculates dynamic effect analysis for credible ATT estimates.

To validate difference-in-differences estimates, run placebo tests, alternative controls, permutation tests, and sensitivity analyses. These robustness checks ensure your causal inference withstands diagnostic scrutiny across staggerated adoption timing.

Yes, this DID workflow supports R, Stata, and Python exemplars. It emphasizes parallel trends, treatment timing, and appropriate standard errors across these environments for credible causal inference.

For staggered adoption designs, TWFE regressions can produce negative weights due to heterogeneous treatment effects over time. The skill helps specify appropriate estimators like Callaway-Sant'Anna or Sun-Abraham to resolve this and yield credible dynamic effects.