>The real procedure that we should teach kids is: collect data, ask questions about it, hypothesize some answers, and only then do a experiment to test your the

>The real procedure that we should teach kids is: collect data, ask questions about it, hypothesize some answers, and only then do a experiment to test your theory. Apparently lots of interesting discoveries have followed the formula of "get drunk with other PhD students, wait for someone to ask "what if...", do experiment, pretend idea wasn't suggested in a bar at conference"