Differences are hidden because (1) differences, even small ones, are used to justify discrimination (2) some feel the need to correct for stereotypes (3) these differences often don't really exist or amount to a small effect size.[0]

In the end, we're talking about distributions of people, and staring at these differences mischaracterizes all but those at the mean.

All that matters is who can pass the test.

[0]: I also encourage you to ask ChatGPT/Grok/Claude "men vs women math performance studies." You'll be shocked to find most studies point to no or small differences.

[1]: Malcom Gladwell wrote a great piece about his experience as a runner that seems appropriate to share https://www.newyorker.com/magazine/1997/05/19/the-sports-tab...

Quite often those differences exist because of systemic or cultural bias that affects the test design. Tests are often validated based off of other tests that showed a difference, but those tests often had a severe sampling bias that showed a group difference where non-existed. It then became an established theory that if you design a test that measures e.g. “emotional intelligence” (whatever that means) and it didn’t show a group difference, it was invalid and had to be adjusted until it did.