lukestein’s Twitter Archive—№ 3,159

@causalinf And you know we probably only even recommend cold compresses for burns because “we’ve always done it that way” or someone did a mean-outcomes comparison with n=20 and non-random assignment…