I made a little extra money participating in psychological experiments when I was a student. But I was puzzled how the researchers could deduce so much about human nature from their select samples of undergraduate students. I know I messed up someone's research project when I was asked to identify sounds by calling one "a Polish L" [written thusly Ł and pronounced something like a W].
Anyway, I now see in the Economist that scientists are finding nonconfirming results when they run classic experiments with non-western, non-rich and well-educated people. One way of finding such populations for their experiments, however, is a crowdsourcing site on the Internet. I guess they can get people with some important differences from US undergraduates, but it's far from a random population sample.