That said I don't think the numbers that Justified used was very representative of the link between homosexuality and pedophilia. There are too many variables to compare the two within in a single survey. Some concerns come to mind when assessing the validity of those numbers.
I like to give the source the benefit of the doubt. D:
But surveys by nature don't really prove anything. There's always the chance that people aren't honest with their answers, that there's jerks who answer a certain way to skew the results, that people just fill them out to win a free car (or NX!), that the proctors mail them out to selective areas, etc.
Assuming it was a voluntary and anonymous survey though, I think it'd lean at least somewhat towards real figures - I don't think many people would be scared of saying something if there was no way of knowing who's answers are who's. I just assume the best case scenario, because there's still a possibility it's true and (if it's not) the people on the other side of the argument can also pull their own biased survey to negate mine.