logoalt Hacker News

mwkaufmatoday at 4:20 PM0 repliesview on HN

So the best way to run a controlled trial is to maximize the external noise and confounding variables?