OK, let's read. Your first paper quotes an optimistic black IQ deficit of 0.33 sigma. Assume strong STEM students are at the 2 sigma level, which is a percentile rank of 97.72% of whites. If whites are 75% of the population, then 1.71% of the population could be a white strong STEM student.

For blacks at 2.33 sigmas, the percentile rank is 99.01% applying to 15% of the population, meaning 0.1485% of the population could be a black strong STEM student.

Those percentages are in an 11.5 ratio, but the actual population ratio is 5. So blacks are about half as likely to be bright as whites. But in US public schools, they must get good grades at about the same rate to avoid "racism" and "disparate impact". The only way to do this is by severely watering down the curriculum.

This is exactly what I was saying about a one-size-fits-all school being unworkable.

The normal distribution is a harsh mistress. The area under the tail drops off very fast, so a small difference in a subgroup translates into overwhelming victory or overwhelming defeat.

Now let's revisit your -0.33 sigma black deficit. For the strong potential STEM students calculated above, the black:white ratio is about 10%. That means we would expect 2 blacks in every university STEM class, and 1 in a typical all-team engineering meeting. In reality the number is far lower. (I've only had two reasonably bright black colleagues ever.) You find equally few blacks in hands-on intellectual jobs like owning a chain of gas stations. So the -0.33 sigma number does not pass the everyday experience test.

Speaking of recommended reading, you could profitably spend a few minutes studying a normal distribution Z table. The normal distribution is a harsh mistress.