StatsPAI v0.2.0

StatsPAI v0.2.0

The Causal Inference & Econometrics Toolkit for Python.

What's New

Post-Estimation (Stata margins / test / lincom)

• sp.margins(result, data=df) — Average Marginal Effects with delta-method SE
• sp.marginsplot(me) — Visualize marginal effects
• sp.test(result, "x1 = x2") — Wald test for linear restrictions
• sp.lincom(result, "x1 + x2") — Linear combinations with inference

Publication Tables

• sp.sumstats(df, vars=[...]) — Summary statistics (Stata tabstat)
• sp.balance_table(df, treat='T', covariates=[...]) — Balance table for matching/DID
• sp.tab(df, 'var1', 'var2') — Cross-tabulation with chi² test (Stata tab, chi2)

Export to Word (.docx)

• sp.outreg2(r1, r2, filename='table.docx') — Multi-model table to Word
• sp.modelsummary(r1, r2, output='table.docx') — Comparison table to Word
• result.to_docx('result.docx') — Any single result to Word
• All tables support: .docx, .xlsx, .tex, .html

Inference

• sp.wild_cluster_bootstrap() — Wild cluster bootstrap for few-cluster inference
• sp.aipw() — Augmented Inverse Probability Weighting

Other Improvements

• PSM now warns about King & Nielsen (2019) bias concerns
• Updated README with Stata/R comparison table
• python-docx added as core dependency

Full Module List (v0.2.0)

Category	Functions
Regression	regress, ivreg, panel
DID	did, callaway_santanna
RD	rdrobust, rdplot
Matching	match (PSM/CEM/Mahalanobis)
Synth	synth
ML Causal	dml, causal_forest
Other	causal_impact, mediate, bartik
Post-estimation	margins, test, lincom
Diagnostics	oster_bounds, mccrary_test
Tables	modelsummary, outreg2, sumstats, balance_table, tab
Inference	wild_cluster_bootstrap, aipw

Install

pip install statspai==0.2.0

280 tests | 78% coverage | Python >= 3.9