you don't need a very random sample, nor do you
need a very large sample. It's got to do with the
fact that our brains largely function the same.
Do you think, if I did a study of political leanings among college undergraduates, that I'd get the same results as a study of the general population because "our brains work the same" ?
This is not a study of political leanings of people. It's a study of the correlation between a characteristic of a person and their political leaning.
If you take 200 college undergraduates, let's say 80% of them are lefties. That means you will have 160 lefties and 40 righties. Then of the 160 lefties 50 believe in BS (31%) and of the 40 righties 16 (40%) believe in BS, you can take those numbers to your statistician and ask him whether that spread is significant.
Just the idea that the correlation is there for these college students, where you might not expect it would be there for college students is a hint that maybe this could hold up for the population in general. At the least it could warrant for a larger investigation.
