lukestein’s Twitter Archive—№ 12,790

@causalinf @SarahHBana @ZOEconomy @dbergstresser @BigOrangeEcon @abiylfoyp “For the panel of individuals who have run in multiple years, we regress completion times on a set of individual fixed effects and their interaction with a linear control for year. We then rank runners by the top of the 95% confidence interval on their interaction coefficients…”