In a way, they are voting with their choices, and all we need is a focus group.
"It should be noted, though, that the largest study to date included nearly 1000 children (9-14 years old) and found no effects.9 This latter study is of particular interest, because it included adolescents, whereas most other stereotype threat studies were carried out with university students. If stereotype threat discourages girls from pursuing math-intensive STEM coursework and careers, its effect should be evident in high-school. The fact that a large and well-designed study could not find any effect, in our opinion, suggests either the effect does not exist or it is unmeasurably small."
edit: and then of course, such dataset should be anonymized and made public, so anyone could see what we are talking about.