xmh

You can run causal inference on tabular data: treatment-effect estimation (ATE/ATT/LATE/ITT), OLS, propensity scores, IV/2SLS, DID/event studies, RDD, robustness checks, falsification tests, heterogeneity analysis, plus data-cleaning advice and identification memos.

Applied researchers, economists, policy analysts, and data scientists conducting policy evaluation or publication-grade empirical research. It supports both English and Chinese users working on causal inference, policy assessment, and robustness testing of tabular datasets.

Upload your tabular dataset and describe your research question. The skill performs diagnostics and cleaning advice, recommends an identification strategy, estimates treatment effects with the appropriate method, and produces robustness checks and falsification tests.

Yes. It natively supports instrumental variables (IV/2SLS), difference-in-differences and event studies, regression discontinuity designs, and propensity score methods, along with heterogeneity analysis and falsification tests for credible causal identification.