Study Design Bleg

I'm interested in the question of whether there are appreciable differences in the quality of parenting received by adopted children and children raised by their biological parents. Obviously this is an empirical question, so before I commit to any empirical investigation I want to be as confident as possible that a potential investigation would really be measuring what I hope to measure.

The standard way to answer questions about the effects of parenting and genes is through twin studies, which compare the achievement and variance thereof between identical twins raised by different families. It's not hard to think of set-ups of such studies which would allow us to test the hypothesis that there are indeed differences in the quality of parenting given to biological and adopted children; however, I'm struggling to think of set-ups that are actually likely to exist in sizeable numbers. Are there really likely to be many cases, for example, of pairs of identical twins where each set of parents raised one member of each set of twins?

The best design that I've so far been able to come up with and subsequently remember long enough to write down is to compare the achievement of children adopted by extended family with that of children adopted by non-relatives. If there is a biological dimension to whatever causes good parenting, we should expect it to be at least somewhat present when raising nephews, nieces, and grandchildren. Unfortunately there's a confound, in that having extended family who are willing and able to adopt your children is itself a fairly good indicator, so we may end up measuring things other than parenting quality.

Any ideas on other study designs or for improvements?