Standardized testing is not biased against underrepresented minorities. The study that UC commissioned found that performance in the tests correlated with college performance and was not biased.

Underrepresented minorities are as a whole not as optimized for the system. Some of that is that cottage industry of essay writers etc, SAT prep courses, etc... but some of it is objectively not being as prepared. Now some of this is also due to growing up in worse environments for academic performance but this is a real difference in performance.

I have interviewed Stanford students. There is a step difference between the Asian/white students and the underrepresented minority students in their performance. The underrepresented minorities were still smart but the Asian/white students behaved like they were a year or two ahead in school.

You could say that standardized testing is somewhat biased against poverty rather than minorities directly. Poverty won’t let you afford extra summer tutoring, SAT prep, etc...

Or the extra opportunities to explore your interests.

Or a quality education k-12 to begin with

Yes

You know you can just type a long comment, right? There’s no need to make a thread where 80% of the comments are just you replying to yourself.

Oh lol I did not see that all of the responses were from the same person.

Hacker News doesn’t let you edit a comment after a period of time.

The best white/Asian kids seemed to be much more experienced with actual computers and computing. Classes taught in schools cannot compensate for this. It’s the difference between learning about programming in school and actual on the job programming or better yet hacking with strong friends for a year. It’s a huge difference in confidence, comfort, and dexterity. It isn’t necessarily a difference in raw ability.

An example: I interviewed an underrepresented minority student from Stanford. He was bright but obviously only understood computing from what was taught in class. So he didn’t do that well in the interview. If it were up to me, I would still have brought him in as an intern. However, the company did not move forward. Under their criteria, they wouldn’t move forward because the student would not have been able to contribute. The company should have changed the criteria to look for raw ability instead of realized ability and to look to nurture the next generation rather than feeding its hiring pool directly.

How did you test for computing skills 'outside' of the class though (you mention it was obvious, I don't see how this can be)? Or is it just your gut-feeling that he wasn't any good at extracurricular computing?

One good step is mentorship, outreach, and internship programs for socioeconomically disadvantaged kids starting from elementary school. The tech giants should just embrace this.

A poor asian family might have a hundredth of the wealth of a rich black family. Race is a terrible proxy for wealth. If wealth is the issue, use wealth.

In the United States, race is a pretty great proxy for wealth.

That's far less true at the statistical margin of elite college admissions. You don't have to think very hard to see why.

It's not even true in general, the variance is enormous.

No it isn't.

> Black Americans on average have a tenth of the wealth of a white family.

And the median?

Median black wealth is 17k and median white wealth is 170k[1]. Were you expecting the difference to be less or more?

[1] https://www.taxpolicycenter.org/fiscal-fact/median-value-wea...

Less, because the 1% are overwhelmingly white. But from [1] it looks like if you take the average then it's actually less than a 10x difference, which is surprising to me. And the difference in income is only 1.5x.

[1]: https://www.visualcapitalist.com/racial-wealth-gap/

I imagine that black wealth is pretty lumpy. Blacks are underrepresented in white collar professions but overrepresented in elite sports and entertainment. There aren’t too many black small business owners. There’s only been a few decades to build wealth so blacks who have some probably made all of it themselves in the last few decades. That’s probably why the average looks better. But that’s all assumptions on my part and I have no sources to back any of this up.

1.5x is a huge difference in income. Imagine a 50% raise.

It is huge - I had originally put the “only” in quotes, but removed them because it’s really not huge compared to 10x.

I do think it makes sense that the wealth difference stems from blacks having so little time to build wealth. The income also is a big factor for sure, but it’s hard to see it causing the 10x gap by itself. OTOH generational wealth is lost really quickly.

conflating standardized testing with the rest of that list seems very counterproductive.