And there's also the problem that you're sampling the environment of people who adopt kids. Conclusions drawn from twin studies don't necessarily hold outside similar families - though it very likely holds for mine :P