[a_gentle_autist]

Some choice quotes:

"So even for a mid-level or lower-level engineer, they'll tell us, "I only want someone who went to Stanford," or someone who has started a company before."

"And most companies in Silicon Valley say, "We hire from these five schools, period," right?"

"If you go to those five schools, the percentage of minorities and the percentage of women is X and Y. Let's say it's 3% and 10%. If those companies hire only 10% of the people they interview, then the number [of minority hires] you get is zero. Right? It's zero."

Absolutely revolting.

[api]

"We hire from these five schools, period," right?"

I'm so glad they brought this up. This is the elephant in the living room, and is a root cause behind a lot of these issues.

The "top ten universities" fetish is by far the most significant mechanism for the perpetuation of aristocracy and aristocratic networks in our society. It's a effectively a feudal system of "knighthood," or at least analogous to that. Those who are admitted to top tier universities are tapped to be at least considered as future members of the nobility.

I'd love to see some real data on employee (and founder) performance vs. where they went to school.

[VLM]

Recruiting only the top 1% nobility is a peculiar power mismatch unless your company really is one of the top 1% nobility of all employers.

What I'm getting at is if you're a bottom 10% marketing experiment startup, demanding to recruit top 1% talent, you're going to get some self selected weirdos. You better check their references and backgrounds to make sure, and its foolish to assume anything about their demographics is going to be normal. There's probably a reason a claimed top 1% applies to a bottom 10% employer and its probably not good. Best case scenario is addiction to extreme risk taking. Worst case, well, its pretty bad, closet full of skeletons or outright fabrication of background or worse.

(edited to add, my example is unfortunately too extreme, but the general rule holds that the wider the mismatch the weirder the applicants are going to be... top 1% vs bottom 10% will as I propose be really weird, yet still, demanding top 1% at a "mere" top 10% employer is still going to be all messed up. Also by "worse" I mean stuff like fired for cause because of sexual harassment and stuff like that, and this explains a lot about the stereotypical bro)

On the other hand, if you want to work at a pretty boring median fortune 500 type company, we have plenty of women and minorities. Unsurprisingly if you demand normal people for normal work you get a normal-ISH distribution of gender and race. Maybe not perfect but not some absolutely laughable 3% stat.

Pro athletic teams demand weird non-median employees... not the median walmart shopper. Unsurprisingly their demographics tend toward the bizarre compared to the general population. Google for NBA racial demographics, its truly weird compared to the general population, as an example.

[a_gentle_autist]

Here's what I've noticed.

There's no correlation.

It doesn't mean anything. It's a position taken out of fear.

At this point, if you've graduated from top ten universities I will ignore your resume, because you won't have the hunger I want from an employee, nor where you have great creative thinking.

[api]

That's been my suspicion for some time too.

"Nobody ever got fired for hiring someone from Stanford."

[jacquesm]

I've hired a fair number of people, college educated (UK top schools), 1 from a top tier US university and a whole scala of other levels of education, all the way to none.

My experience (which is very limited) was that the people that were less educated had more drive because they were given a chance but the people that had a degree of education that was generally perceived as higher were more productive even without that level of drive and the quality of their output was generally higher.

This is probably not surprising but with very few exceptions that seemed to be the rule (and those exceptions were totally off the scale).

[api]

So you'd say that the median performance is higher from the top schools, but that the difference appears less pronounced at the extreme ends of the curve?

Basically this would mean that an average person from Stanford would probably be better than an average person from University of Nowheresville, but that an exceptional person from the latter might be as good as an exceptional person from the former...?

If true this would account for a tendency to try to recruit from top schools, since the odds of getting a better candidate might overall be higher. But it doesn't change the overall social implications much.

[jacquesm]

No, an exceptional person without any formal education can blow a person with a formal education clear out of the water both in drive and in productivity and quality.

But that's 'exceptional' for you, it is an exception, I've only encountered one such person to date.

[pimlottc]

> I've hired a fair number of people, college educated (UK top schools), 1 from a top tier US university and a whole scala of other levels of education, all the way to none.

Did you mean to say "scale" here?

[jacquesm]

No, I meant scala. It's not just a programming language :)

scala(Noun)

Ladder; sequence.

http://www.definitions.net/definition/scala

[dabockster]

However, the unfortunate truth is that no one is willing to risk getting fired over not hiring the very best. And often, this means that only the top tier schools have a chance.

[jacquesm]

The laws of limited supply are your friend here. They may want to hire only from the top tier schools but since these are typically only available in certain numbers and usually already gainfully employed this becomes less of a problem because they probably can't.

[humanrebar]

> If you go to those five schools, the percentage of minorities and the percentage of women is X and Y. Let's say it's 3% and 10%.

I am also miffed at that kind of elitism and parochialism.

That being said, both of those numbers are very low for general enrollment at top universities. MIT is a little over half male and about a third white. Even if you don't count Asians (I guess they're not minorities?), 3% and 10% are extremely lowball estimates. Or are we talking only about students with software-related majors? Does anyone have stats that describe the demographic breakdown for computer science majors?

Source: http://web.mit.edu/ir/pop/students/diversity.html

[tzs]

Asians are essentially the 21st century's Jews as far as universities go. The universities want some, because they bring up the academic stats, but they don't want too many of them so they tweak their admissions processes to not let too many in. They just can't be as honest about it as they could with Jews in, say, the 1930s when they actually had explicit public quotas.

From 1990 to the present, the number of Asian college students nearly doubled, but if you look at most of the top universities, Asian percentages have been pretty much flat, with the exception of MIT and Caltech. At Caltech, Asians are now the biggest group, at 47%. Whites are only 33%. Next comes Hispanic/Latino at 11%, multi-race (not Hispanic/Latino) at 7%, and Black at 2%. MIT incoming class of 2018 is around 30% Asian (50% White, 14% Hispanic/Latino, 11% Black) [1]. (45% women at MIT, 40% at Caltech).

Caltech's distribution is probably closest to an accurate representation of the demographics of top STEM students coming out of high school, because Caltech does not take into account race or gender when making admission decisions. There's still some distortion, though, for a couple of reasons.

First, Caltech does actively seek out strong minority and female STEM students and try hard to persuade them to apply, and they have run projects that have provided summer science programs for high school students in minority districts to give them a boost. Once someone applies, however, race and gender are not considered for the admission decision. In 2010, according to numbers I saw, this resulted in 106 Black applicants (out of 4859 total applicants). 19 were accepted (out of 610 total acceptances). Only 6 of those decided to attend.

Second, the Asian numbers at Caltech are probably boosted a bit by the de facto Asian quotas at the Ivy League schools, UC Berkeley, Stanford, and other top schools (other than MIT). There are probably several Asians at Caltech who would have picked one of those other schools if they hadn't been shut out because of their race. MIT might get a boost from this factor, too.

There's probably also a bit of an Asian boost at Caltech because of location. There is a noticeable skew in Caltech's student body toward people who come from the West side of the country, which is also where Asians are more concentrated.

[1] http://mitadmissions.org/apply/process/profile

[potatolicious]

> "Even if you don't count Asians (I guess they're not minorities?)"

We are, and we have our own problems with race and tech (see: representation of Asians overall vs. representation of Asians in leadership positions), but the Asian-American experience is fundamentally different from the Black- or Hispanic-American experience.

We are also frequently used as bludgeons against other races. As if the disparity between Asian enrollment in schools and Black enrollment comes simply down to working hard enough. It's important to note that some of this "Asians as bludgeons against other races" comes from Asian individuals.

Asians in America do not fit into a simplistic race narrative. We are not "as good as" whites, but neither do we have many of the same problems that face Blacks and Hispanics.

There was an article I read about the nature of being Asian in tech, and one quote stuck out to me: "we have a problem, and we are a problem, at the same time".

The presence of Asians should not be mistaken as the lack of a race and diversity problem.

[theorique]

We are not "as good as" whites

Well, in some key ways (at least for tech) Asians are better - e.g. a +0.5 SD average IQ boost really helps for the cognitive and analytical challenges that are commonly encountered in the tech/programming worlds.

[santaclaus]

> Even if you don't count Asians (I guess they're not minorities?)

Not according to the NSF, strangely enough -- people of Asian decent are not classified as an underrepresented group within STEM fields. Pick another US .gov agency and it might be different.

[seanflyon]

It's worth noting that "minority" and "underrepresented group" mean different things. I think you are right to talk about underrepresented groups instead of minorities because under-representation is at the heart of the issue.

[Uehreka]

> "If you go to those five schools, the percentage of minorities and the percentage of women is X and Y. Let's say it's 3% and 10%. If those companies hire only 10% of the people they interview, then the number [of minority hires] you get is zero. Right? It's zero."

I agree in principle, but this math seems off... Sure you only hire 10% of the people you interview, but if all the people you hire come from these 5 schools, then you should still have 3% minorities and 10% women (since 100% of your employees are from these schools).

3% and 10% are plenty revolting, there's no need for them to fudge the numbers.

[itchyouch]

The other bias that shows up is the different requirements portrayed to people based on their race. There was a recent article about how this was done in Lousiana to keep an apartment community predominantly white.

I see my company pulling a similar stunt. Employees are strictly hired through a recruiting firm and only after passing the interview do we bother to have them actually submit an actual application where the gender & ethnicity stats are collected. I suppose this enables us to be able to say, "we have hired 100% of all black/hispanic/asian/etc people that have ever applied to our corporation. There is no way that we are racist at all. We can't help it that 99.9999% of applicants ever, were white."

[mcguire]

Human beings only come in discrete quanta.

[guard-of-terra]

"We hire from these five schools, period,"

This is a strange thing for them to do. I know on occasion that they do hire immigrants; who certainly did not go to "these five schools".

[dustingetz]

What? Can anybody comment if it is true that many SV companies only hire from a handful of elite schools?

[andymcsherry]

Well, I can tell you that not only did I not graduate from an elite school, I didn't even graduate from college. I've never had a problem finding work at startups. Very rarely does anyone care -- it's always about whether or not you have the necessary skills.

[santaclaus]

Maybe not intentionally, but if your workforce is mostly MIT/Stanford grads and part of your recruiting strategy involves sending employees to their Alma Maters and leveraging their social networks, there is a bit of positive feedback in play biasing recruitment towards MIT/Stanford. Totally anecdotal, but there was no Apple recruitment presence at the school I went to until a few students were hired into Apple.

[aantix]

I work as a consultant in SV doing Ruby on Rails.

I've done 100+ interviews and never once did the school come in as a prerequisite.

The teams are generally small. We make the coding exercises intense. You will be asked to code onsite. At the end of the day, we just want a guy that we're not constantly cleaning up after.

[NearAP]

A few companies I know only do college recruiting from specific schools (the elite schools). The Director of College Recruiting for one of those companies told me - we only go to those schools for recruiting events; however, anybody from any school in the US can apply for any vacancy listed on our website; the person will simply have a longer/larger pipeline to go through.

I should add that the companies that I'm referring to are NOT startups.

[seanflyon]

It is not true.