did_causal_analysis

Difference-in-differences estimates causal effects by comparing pre- and post-treatment observations across groups. This Skill calculates DiD estimates, p-values, and confidence intervals for policy evaluation using your panel data.

Yes, causal inference supports multivariate heterogeneous DiD with interaction terms to estimate collective treatment effects across multiple features. It returns per-feature estimates with p-values and confidence intervals for robust interpretation.

For small sample sizes, use the univariate DiD fallback to estimate treatment effects. This approach provides robust per-feature estimates with standard errors and p-values without requiring the data volume needed for multivariate models.

Structure panel data with a 'Period' column indicating baseline or treatment, feature columns for groups, and a numeric outcome column. The Skill supports both intensive sparse margins and extensive complete panels for proper causal analysis.

Yes, difference-in-differences is applicable to A/B testing, event studies, and policy evaluation. It identifies causal effects of interventions by analyzing pre- and post-treatment observations across control and treatment groups.

The DiD analysis provides effect estimates, standard errors, p-values, and confidence intervals. These transparent statistical outputs allow you to test significance and interpret treatment effects across groups and time periods.