Hacker News new | ask | show | jobs
by bombcar 1462 days ago
Way WAY too many "studies" are done this way - the famous "50% can't afford a $500 unexpected expense" study for example is a survey asking people if they could spend $500 without worry.

It's cheaper, but it's not better. True investigations watch results over time and compare actual items (that still have to be determined to be significant).