Panel Data Multiverse Lab

A real longitudinal wage panel meets the multiverse. Pick an effect, then watch the whole universe of estimates open up as you change estimator, controls, sample and inference — and learn why the answer moves. Data: NLSY · 545 men · 1980–1987. Every number is verified against R's plm.

🎯 The big idea

One dataset, one question, many defensible analyses. The spread of results across reasonable analytic choices is the “garden of forking paths.” Here you build that garden by hand on real wage data and see which conclusions are robust and which are an artifact of a single specification.

Loading real panel data…