The “twin method” has been found to be unreliable, as have investigations based on adopted children, probably because the sample sizes were always quite small and the evidence was prone to subjective error. There is a lot of anecdotal evidence about twins, and superficially it is very convincing, but deeper investigation based on the details of genetic markers revealed that many of the links between specific characteristics and genetic structures hypothesised on the basis of twin studies turned out to be wrong.
The current work on genetics is showing that DNA works in a much more complex way than anyone imagined 20 years ago. Individual genes are not good predictors of anything except some really basic physical characteristics like eye colour and some very rare diseases (eg Huntington’s is caused by a mistake in a single gene). Instead, it is the interactions between genes and also the way genes can be “switched off” that seem to be linked to more common conditions and traits. Even then, there are often several different possible combinations of genes that produce apparently similar results.
There are environmental factors, including such unexpected things as extreme trauma, that can cause genes to be switched on or off, and these changes in the genome make it actually impossible to distinguish between “nurture” and “nature” — because the true answer is that the cause is both, at a fundamental level.
We seem to be just as far from a clean genetic explanation of differences between humans as we were 20 years ago. The more we learn, the more puzzles we find.