The short version is: think about how heritable something like "number of arms" would appear in a study like this. It would appear that variance in arm count is 100% environmental, since the only people who don't have two arms lost one in an accident of some kind. Arm count is therefore not controlled by genes!
any causations of this type (gene causes something detectable by parent, and parental reaction causes improved executive function) is consistent with the study/data.
but i find it hard to believe that environmental effects of any sort could explain 99% of the variance that we see here, and I don't know how even a whole genome analysis could rule out GxE interactions. how exactly is that done?
what can be done is to use arguments to determine which explanation is better.
i think the most important argument is this one:
we can explain how culture/memes (triggered by some detectable but unimportant effect of genes) would be able to affect intelligence, cause ADHD, and many other things. that is not a problem.
no one has a good explanation of how genes have complex and lasting effects on human minds and personalities.
of course for both points further elaboration is needed. but anyway, no one has made a strong case the other way.
this may be of interest: http://settingtheworldtorights.com/node/269
edit: regarding 99%, i don't think it's a problem. why wouldn't highly adapted memes be able to cause effects reliably?