Claude Code 31: Apple-to-Apple Audit of Six Callaway and Sant'Anna packages

Six Packages, Same Estimator, Same Specifications, Same Dataset, Different Numbers!!! :(

This is part of a long-running series on Claude Code which began back in mid-December. You can find all of them here. My interest is consistently more on the end of “using Claude Code for practical empirical research” defined as the type of quantitative (mostly causality) research you see these days in the social sciences which focuses on datasets sto…