"Stereotype threat"
is junk science.
You may have heard about the replication crisis in psychology... Well, the studies cited as proof for the "stereotype threat" theory are squarely among those that can't be replicated.
All things considered, it looks like publication bias is the only reason the stereotype threat theory ever got as far as it did:
The academic left looooves the "stereotype threat" theory for ideological reasons, and so over the decades, leftist-dominated psychology departments have performed many studies attempting to proof it's true.
But due to small sample sizes, shaky methodology, and the inherent fuzziness of social science, pretty much all of those studies measured mostly just random noise.
The ones where the random-noise results happened to coincide with the
wanted results, they published them.
The ones where the random-noise results happened to go in the other direction, they canceled the studies and didn't publish the results.
Then they did meta studies which smugly proclaimed: "Wow look, all these studies finding evidence of stereotype threat! No need to even evaluate the quality of the individual studies, there's so many of them!"
And that's how stereotype threat became a "known fact" among the academic and journalistic left.
Only in recent years, with the large-scale efforts to test old results for replication (and thereby uncovering the replication crisis), has it turned out that when those "stereotype threat" studies are re-done using the same sample sizes and methodology, the results are randomly for or against the theory. Whereas if they're redone with larger sample sizes and more robust methodology, they reliably debunk the theory.