You might think an error made with total confidence would be the hardest to shake. It's the opposite: in 2001, Brady Butterfield and Janet Metcalfe showed that wrong answers given with high confidence were the most likely to be fixed on a later test, once the right answer had been revealed. The surprise of being wrong seems to grab attention and lock in the correction. This “hypercorrection effect” shows up in children too.
Remembering is not replaying a recording — it is rebuilding. Each time you recall a memory, it becomes temporarily malleable and can be altered before being stored again. That is why two people can sincerely remember the same event… differently.
In 2003, Johnson and Goldstein compared European countries. Where you must tick a box to become a donor, consent rates ranged from 4% to 28%. Where you're a donor by default unless you opt out, they topped 85%. People aren't more generous on one side; most simply stick with whatever option was chosen for them.