Regression analyst · preview

For advanced, frontier causal inference.

A specialist expert for serious econometric modelling — beyond a default regression. It chooses and justifies the right estimator for your question (staggered difference-in-differences, event studies, honest DiD, synthetic control, IV / marginal treatment effects, shift-share / Bartik designs) from a validated methods corpus — the methodological papers behind each design, paired with the R package that implements it.

Live now: it grounds every recommendation in that corpus (citing the source and naming the package), and runs the deployed panel estimators — event study, difference-in-differences, honest DiD and two-way fixed effects — on real data, with clustered standard errors and an honest account of the identifying assumptions and where they break.

Coming with the dedicated instance: the full frontier estimators run in R (Callaway–Sant'Anna group-time ATTs, synthetic control, KSS leave-out variance components on matched worker×firm data) — built into the analysis file together with the Data cleaner.

Launch the Regression analyst →

Want a private one now? Clone your own agent on kojai.no →