If you care about diversity, don't just hire from the same five schools

Agreed--I was accepted everywhere I applied save one and I ended up going to the University of Maine. I've never hurt for work and I paid off my student loans by the age of 25. Google sniffs around now (aside: recruiters, even if it's Google maybe you shouldn't assume I want to talk to you and tell me to sign up for a phone call to talk to your sainted ass?) and I have no interest, but the first time I interviewed, when a guy at Google gave me noticeable shade for my filthy state school degree? They could've made a decent bit of money off of me and now that door is likely closed.

Can you really expect any different from a company founded from within the halls of Stanford, and then staffed primarily from there?

On the other hand, I bet if you sat down and ran the numbers, recruiting from state schools isn't productive. For better or worse, school selectivity is tied with rankings -- for every Anthony topping the charts of state school is a graduating class of Anthonys at the elite schools. Cherry picking your one Anthony per school is likely harder than showing up to Stanford where they picked cherries 4 years ago.

Moreover, Google recruiters are likely already flooded with new grad applicants from non-elite schools, while every year GoogFaceAmaple is pushing forward their recruiting funnel earlier than the competition. Eventually, I imagine they'll simply interview candidates the summer between receiving their college acceptance letter and fall semester freshman year.

why stop here though, couldn't you just extend this argument to people who go community colleges or online schools or even people don't go any school ?

State schools have people that got into MIT, but didn't go for financial reasons or family reasons. Nobody at RPI got into MIT because if they did they would have gone to MIT (maybe some rare exception exists, but this is generally true).

School admission is obviously not a perfect indicator of ability, but I suspect while the average and median may be higher (and more tightly distributed) at a school like RPI, a state school will have a lot more outliers.

Everyone is fighting for students from the super selective schools and state schools are largely ignored - students either need referrals to get interviews or jump through a ton of hacker rank like hoops.

This is even ignoring the 'diversity' angle - a lot of really great people are missed because it's hard to break in without knowing someone or having the right credentials.

Don't they have a referral bonus scheme pretty much exactly for this?

I was puzzled why the recruiter wanted to know my SAT score. He also wanted to know the college and high school I went to and my GRE score and my undergrad GPA ... Keep in mind that I have a BS, MS, Phd in three different engineering fields and over a decade of work experience at well know companies in the Bay area. For a 40 something engineer it seemed odd he would ask me information from two decades ago but I went along assuming he was filling in some boxes for HR.

The first assignment in the Probabilistic Graphical Models class was to predict the likelihood that a candidate would be a good fit for a job given prior Bayesian probabilities for ivy league school attended, SAT score, GRE score, all forming nodes of a decision tree.

I looked at the problem assignment and the recruiters questions and realized what was going on. Some Stanford grad at Google had convinced HR that they needed an objective way to evaluate hundreds of candidates and what better way than to use the same methods that they had learnt from Daphne Koller at Stanford.

I quit the PGM class 'cause it seemed like a tool I would never use. I decline interviews at Google since I don't have the right Bayesian probabilities that would trigger a positive outcome for their PGM model.

Compare that to the students I interviewed for internships from top 5 engineering schools. Among them was one of the best candidates I'd ever encountered, regardless of experience level. He came up with a unique (and quite good) solution to a problem I'd been using as an interview question for years. As someone who had done fairly well in state school, I was floored by how much more well-prepared these students were than I had been.

My story is anecdotal, and there are many factors that can explain what I perceived. But I suspect the top companies have figured out that they get best results in hiring based on a) recent grads from top schools and b) great resumes from experienced devs who maybe didn't start out at top schools & top companies.

On the plus side from a diversity standpoint, the students I interviewed from these top schools were both male and female, and from a wider variety of ethnic backgrounds than I usually see in the hiring pipeline.

I interviewed with them recently and I didn’t get any condescending comments about my SUNY education like I did a few years ago — from a recruiter who was reaching outbound to me!

And how about people who couldn't get into an elite school at 17, but end up learning a lot at the state school? It seems odd to be so preoccupied by what someone's capability was when they were in high school, rather than what it is now.

My experience is that in any school / college / organisation, there is a distribution of smart / switched on / less switched on / absolute idiots. The difference between the better universities and less prestigious ones isn't that every student is excellent or bad (I have seen enough falling in the last category above in prestigious universities), just that there is a higher concentration of better students.

So it does make sense for companies to target these schools / universities, it's a matter of efficiency for their marketing efforts.

It doesn't make sense for companies to make it hard for non target schools as they are cutting themselves from lots of potentially excellent candidates (I suspect certain companies to automatically reject applications on their websites if the university isn't one of a list).

Dropping out of that place was still one of the best decisions I've ever made though.

At this point I tell them not to bother if I get contacted or I just ignore the recruiter. It's effectively a pretty big waste of time 98% of the time and they're just making some kind of internal metric so that positions can be claimed to have been offered fairly.

I think what really bothers me about this process in general, and it applies to a great many companies, is that they contacted me. In that they sought me out. I didn't apply. So a little courtesy on that front would be helpful.

I don't understand this. If you're matching the diversity of the market, that would seem to me that your recruiting practices are unbiased, which is good, right?

I've always assumed that open source would be one of the places to go for such telemetry, but have never known where contributing would actually furnish that kind of benefit. It would be good to know if there were projects that are actually fit that criteria, instead of merely being coincidentally worked on by a lot of employees.

Doing my undergrad and PhD at an average large state school, I was really lucky to get an internship (and now job), essentially by word of mouth and my PhD advisor.

Now that I'm doing a part-time Masters at an elite-ish (top 10) school, I'm realizing how amazing career services can be. They send out weekly emails saying:

"James Dimon is coming to speak." "Dir. of Engineering from Exxon will be taking questions." "Bring your resume and chat with recruiters." "Women's Society will be hosting Barclays tonight."

My state school had nothing. All of the leg work was done by individual students.

https://www.joinhandshake.com/

It's founded by 3 graduates of Michigan Tech University, a small but well-regarded school that even other Michiganders think is far north. Their goal is to level the playing field for internships and first jobs.

It was great to meet people so sincere about solving this problem.

I'm guessing that "Career services are shit." is more accurate.

We're attached to the marketing team (~80%+ women), but our company is in a very unpopular industry (plantiff's law).

Then the college of natural sciences took over. I'm glad I was gone; from what I hear, it sucks.

The underlying assumption is definitely an assumption, but it seems like a good zeroth-order assumption to me.

You can tell America's history in terms of decreasing inequality of opportunity. We started with only well-off white men having the vote; black people were chattel and women were pretty close. We have sporadically worked ever since to repair that. One civil war, assorted constitutional amendments, and 8 civil rights acts so far. We try something, see if that's enough, and then try again when it isn't.

Each time people hope that we've done the last thing. The post-Civil War period, the Reconstruction, was supposed to fix things. But then we had the Lilly White Movement. And Jim Crow. And the Nadir, which involved violent ethnic cleansing across the US (mostly but not entirely anti-black).

We certainly aren't there yet. Marital rape wasn't fully outlawed in the US until 1992. Redlining wasn't illegal until the 70s, and the effects on wealth distribution persist to this day. Men and women reached parity in law and medicine degrees a decade ago, but as far as people practicing it's still 2:1. Employment discrimination, as documented by a number of studies, still persists.

So yeah, if some particular solution doesn't fix the problem, then people will definitely keep pushing. Because that's our 200-year history of fighting discrimination: it keeps not getting fixed.

You have more diversity of thought and behavior across social classes or countries: a person coming from Chile to work in the USA will add more perspective than most kind of American born person. A person raised in a rural area will add more to a company comprised of city dwellers than most other city person.

Continue with projects like codejam, though ideally come up with one that is less reliant on obscure math (or OEIS) and more reliant on problem solving.

Also: protip for people in the South or Midwest who are open to relocation: change your online profiles to say you're in the new location.

    > How could companies reach the great students at unknown schools systematically?

Maybe their talents are best directed at small-ish companies where they can contribute, grow, and do interesting stuff outside of monstrously large companies with cut-throat competition?

Having ALL the "top" candidates vacuumed up by the "top" employers just seems wrong to me. I feel it's more healthy to have the best students spread out all over the place rather than concentrated in the top 5 or 6 employers in Silicon Valley. After all, there are more employers than there are colleges.

Google and Microsoft makes about $200k/year profit per employee, Facebook 4x that. The problem, I suspect, is that you probably don't net anywhere near 1-2 hires per couple of events at these schools.

https://www.recode.net/2017/8/4/16090758/facebook-google-pro...

If only a tech company could figure out a way to use computers to scale up processes so that they work for more people.

With our model, we free up companies from having to worry about exactly what you described because we incur the vetting.

I find fizzbuzz and other coding challenges much more effective in hiring.

It's hard to do but showing actual code from your codebase with an actual real bug is a much better indicator of whether someone will be successful than anything else.

I find myself wondering at your definition of "2nd tier school". If you mean community colleges or something, I'd understand. But there's plenty of "tier 1 schools" for learning computer science outside of the "same five schools" in question.

I don't see a list of the five schools in question from a quick skim, but is the University of Michigan "second tier"? University of Waterloo? Any of dozens or possibly hundreds of Division I schools with high quality programs?

Let's be honest... running a fairly good undergraduate computer science program is not rocket science. It's not that hard. For all the sound and fury of the industry, it has visibly not shifted all that much in curriculum in the 20 years since I took it. I know, I look over the intern's shoulders sometimes when they do homework here and I could almost hand them my own homework solutions from 1998 as a cheat sheet. [1]

The very elitism the article is deploring is on full display when people seem to assume it's these top five schools, then a country full of drooling morons. That is not in fact how it works. It's not even close to how it works. It is offensively wrong.

[1]: This is mostly a good thing, not a complaint. The curriculum should be stable. Bits of it need to be updated here and there, but the whole is solid. AI really needs an update, though; it was long in the tooth when I took it in 1999 or so and it hasn't gotten much better at the local schools. The whole "search the solution space" is certainly a bare minimum to understand the field but almost everything has gone in a very different direction since then.

So your implication is that the ratio of "competent" : "completely incompetent" for students coming from these schools is roughly... 2:100?

I know you can't possibly mean that. But your language clearly implies that ratio. The only question is why you're making such a bizarre implication.

I often find that those who are so keen on “hiring broadly” have never worked in a company that hires primarily from Podunk State

Was the original article suggesting that your company hire "primarily" from Podunk State? Or simply that you probably don't want to hire near-exclusively from the Top 5?

The key is not to focus on the schools (rather than to focus extra on state schools. Instead, develop methods for identifying talent based on its own merits, regardless of where it is or what university it went to.

Which is exactly what the original article was advocating - a "completely blind, skills-first approach". In particular, it was definitely not arguing that one should "focus extra" on state schools.

I doubt it would be true over the entire university CS populations. If the only thing you know about two applicants is their university, then the student from a top tier university will on average perform much better than the student from a low tier university.

When I myself was hired by Vervoe, I didn't submit any resume, because as our CEO says, resumes are documents about past and don't necessarily show how the candidate will perform now and in future. I only submitted the skill assessing questions. My school (I have none), age, sex, ethnicity, location didn't matter.

I am glad that articles and comments like this exist, shows we're not alone on our mission to show that diversity works better :).

The official term for this is "distance traveled", and there's a bit more about it here: https://medium.com/kapor-the-bridge/dear-investors-so-you-wa...

Having grown up quite poor in NYC, I was always dismayed by not being able to take advantage of affirmative action programs...but getting selective negative points is just plain unfair.

like "oh, one of your parents was a student here? here, you're in".

Harvard is one-third legacy[1].

> at five Ivy League schools, Dartmouth, Princeton, Yale, Penn and Brown, as well as 33 other colleges, there are more students from families in the top one percent than from the entire bottom 60 percent.

stringent requirements, sure, but certainly not the right kind.

[1]: https://www.cnbc.com/2017/09/06/harvards-incoming-class-is-o...

If I can hire cheaper from candidates not in the top 5 the money could be used to recruit better/same talent at lower costs plus generally a longer employment term.

However, I don't really agree that this is a "diversity" problem, except perhaps weakly so in that it optimizes against people who don't want to study algorithms for interviews. But interviews by definition optimize against some subset of the general population; unless you're defining diversity so loosely that it can be satisfied by the sets `{studies algorithms for interviews}` and `{doesn't study algorithms for interviews}` (and in that case, how are they different)? That's sort of like saying a tech company shouldn't optimize its interviews for people who like to watch baseball versus people who don't - they absolutely shouldn't do that, but that's meaningless as a diversity metric, in my opinion.

How are you defining diversity?

If enough political pressure is placed on SV over class concerns like this, where would the measurement come from? It took statistical evidence to get movement on the more biological concerns, but I'm not sure anecdote alone could have done that. Either way, the result will be little more than a website and some youtube videos.

For example, your statement "A company like Apple has little use for someone who has only ever worked on IT-focused Java development.".

Actually, Apple has a need for people like these. Their whole web/it/retail infrastructure is a Java backend ( i believe webobjects moved to Java, then was deprecated as an outside project, but kept for internal use, my knowledge is dated on this). Take a look at their jobs site[1]

Now my point wasn't about apple and java. But our own biases that we bring, at every level. We all have biases, but most direct way to effect biases in hiring, is to close your pool down. You need a really wide opening at the beginning of your sourcing funnel to have diversity ( and i mean diversity at every level ).

I work at one of the "big 4" and went to an ivy, but guess what, some of the smartest people I have ever met, had to go to schools based on what their parents could afford. Thats a HUGE selection bias.

So yes, widen your funnel, and recognize your biases ( that we all have, and that basically boil down to having less knowledge ).

[1] https://jobs.apple.com/us/search?job=112907897&openJobId=112...

Secondly, a developer that masters and shows proficiency in any complex software ecosystem, including Java and web middleware, is going to have transferable skills valuable to any software company.

In an industry where 60% of the workforce hasn't learned the field through a traditional degree[1], and where a nontrivial fraction of these are self taught, I think that's a pretty bad assumption.

Besides, this argument would hold more water if companies actually tested for this stuff. Most companies seem to be interested in testing algorithms and puzzles, which ultimately aren't actually that useful (sure, algorithms are useful, but not the culture of hyperoptimizing-by-big-O; when n is rarely huge O(f(n)) doesn't matter). Yeah, for an AI job they'll test AI as well, but large companies continue to test with an overwhelming bias for this. I don't think they get to say they want to choose the skillset of their pool by choosing colleges if they do that.

Also, all the large companies (Google, MS, Facebook, etc) hire-first-ask-questions-later; they don't hire college grads for a specific position, they just hire them and work out the position later.

[1]: https://techcrunch.com/2016/01/12/unlocking-trapped-engineer...

If you are "someone who has only ever worked on IT-focused Java development" that apparently we don't want at Apple, my email is in my profile.

Reputable professors are not necessarily the best teachers... Especially at the undergrad level.

Apple just doesn't want to deal with 2-4 months of ramp-up when they can have someone with expertise in a specific subject. Wealthy institutions see that, and teach certain frameworks as part of their graduates' education. Why else would you see core courses in languages like Objective C?

Any graduate can learn it fairly quickly, but a common theme among employers these days is that they will invest nothing in employees.

This might be true for graduate level programs, but between decent state schools and Ivies, the difference, so I have found having been in both "types" of universities, in the curriculum is far, far less than you would lead us to believe. From having met and interviewed a number of people from both, the talent of the individual seems to shine through regardless of the school they attended (if it was at least reasonably decent, i.e. most research-oriented state schools).

I find it odd that you have chosen to respond to an article on diversity with comments that display a knee-jerk bias. Also "IT-focused Java development" is not even a real term. I can only imagine you mean Enterprise Java Development.

I can assure you a quality engineer with great problem solving skills, CS fundamentals and real-world experience working on large important projects with real deadlines and bottom line impact can be of great use to a company like Apple.

No, they don't. At least, not at the undergraduate level.

I went to an unknown undergraduate CS program. Later, I did a PhD at one of the top schools in the country. For my career, I have worked with graduates from all of the elite programs. There's no difference in training. The big-name schools may attract "better" high-school students, but even that advantage is overrated.

Undergraduate programs are largely identical. It's only when you get to graduate school that the resources of a top-tier program start to become relevant.

Yes.

> If the latter, they will have absolutely no problem getting hired.

The whole point of the article is they won't.

What if I told you that diversity hires were paid significantly less? Do you really think all that Silicon Valley PR is out of the goodness of the heart?