lukestein’s Twitter Archive—№ 7,336

@JedKolko I love *your* per-cap regressions but—in line with a branching thread with @paulgp the other day—I’m not clear how we can learn *anything* from R² on scale-variant cross-sectional regressions (even with log transform). Does the 0.81 mean anything to you? @lukestein/1246113753130627073