I don't know how much "(White or Good) and (Black or Bad)" being psychologically-ingrained categories has to do with race, though.
Some things early homo sapiens knew about, that are white: rice, fatty tissue, soap, and safe sunny environments.
Some things they knew about that are black: mold spots, dried blood, charred flesh, and dangerous moonlit environments.
(Same for red with "sweet berries, fresh blood, flushed skin, and fire"; green with "medicinal plants, unsafe still water, vomit"; etc.)
I bet there wouldn't be nearly as clear an association if the colors were "vaguely dark beige" and "vaguely light brown".
Interesting point. Just to clarify, in the original study you were to sort actual images of faces. So when I say white and black, I don't mean they were sorting things such as "rice" to the "white" bin.
My point (as describe in a above comment) was to illustrate the language used when people describe men and female in their goals. Some concept has become so glued together, than we barely can recognize them outside of their decided roles.
I personally think that coeducational studies are more productive, but describing why is hard because it require a kind of bluntness that is unacceptable in current culture.
I'm curious, if you wanted to make a similar point as the article, but imply that coeducational studies are more productive, how would you do so?