rich r,
this problem is so well known that I doubt that the error of not using a same sex dizygotic twin control has been made in any study within the last couple of decades. If this error would have been made in the above article the article wouldn't have made it into a prestigious journal.

En,
that's a real problem (that people who share more of their genes often also live in a more similar environment) and, unfortunately it is not even completely solved by testing twins reared apart for mainly two reasons:
(1) Even twins reared apart have shared the same environment for a period of time (at least for nine months, mostly longer, for the placement into foster care happens rarely directly after birth).
(2) There is nearly universally found what is termed 'selective placement'. Foster children are for very good reasons not placed into families at random. And it is known that there is a bias in the persons responsible for foster placement to choose a foster family that is in some respects more similar to the biological mother than can be expected by chance alone ('Look, little Dana baby comes from a very good religious family, let's her rather give to the Smith's than to the Peterson's').

The most extreme case: The notorious Burt had among his 'monzygotic twins reared apart' twins that were, after the death of the biological parents, placed into the homes of close relatives living in the same village a few yards apart.

Wolfgang