True. Difficult but even if there's some bias you can still get useful data I think. There are things that do not fit well with the basic evolutionary psychology model such as homosexuality. You'd think, using the EP model, that genes corresponding to that would die out however there are a few gay guys out there which suggests you might look for other explanations. Just to say it's not the case that you can find experimental evidence for all EP type hypotheses. Some fit, some don't.