"This kind" being the research we've been discussing, which is about socialisation and toy preference. Finger length and other easily-measured physical differences haven't been part of it.
It's reported in all of the child studies discussed in TFA. Fault the article if you want but "hormones" is in the title, and the whole point of the entire line of research is that it's implausible that prenatal hormone levels could be affected by cultural biases.