Hacker News new | ask | show | jobs
by apl 5242 days ago
Nitpick: You'd have to test the difference between before and after at a certain level of significance, not just establish estimates for before and after. Different things.