t1mmyb’s Twitter Archive—№ 37,000

@JulianDeverell only if you believe every study, the badly-designed as well as the well-designed. Not a reason to make policy from anecdotes