There's some fascinating stuff laid out in this article about how social expectations on test performance appear to have a very real effect on those who are taking the test, sometimes in ways that I (from my own demographic cubbyholes) wouldn't expect. For example, women told to picture themselves as a "stereotypical male" were later able to perform much better on mental rotation / spatial visualization tests than those who were not.
We tend to attribute to human hardware a lot of effects that are due to human software.
“Picture yourself as a stereotypical male” | MIT Admissions
Sep 3, 2015 —
There is empirical evidence to support the idea that males have a higher capacity for spatial reasoning than females. A large-scale 1995 meta-analysis found that on average, men outperform women in a cluster of tests related to spatial ability by nearly a full standard deviation, and in attempt to explain this, researchers have hypothesized about the impact of testosterone and…
Gotta love the comments.
There’s much more along these lines in Cordelia Fine’s excellent book Delusions Of Gender