"There are no Black/African-American GitHubbers in management positions, which is unacceptable."

Why is it unacceptable? Just by virtue of the number of different kinds of people -- race, gender, etc there's no way to have 'one of each' for every position for maximum diversity. This sort of thing is so creepy because it treats people like collectible figures ("Oh yeah? Well I added a black transgendered person to our team today! I have way more diversity points than you!") instead of actual people you judge on their merits. How is a white person supposed to feel about applying to Github management position after reading "Its unacceptable we don't have a black manager"?

I'm not white and I don't want to be treated as some statistic you can show off on your diversity report card. Thats really demeaning and insulting.

Because the percentage of black GitHubbers in management positions is far below e.g. the percentage of e.g. black people with CS degrees in the US, indicating a probable bias in their hiring, retention or promotion pipelines that they're announcing they find concerning.

You're making it about individuals, not GitHub. They're comparing statistics.

Was that percentage equally high in 1990? Fresh grads are not qualified for management.

We can determine bias simply by comparing one company's numbers to national trends?

I said probable bias, and yes.

As you reduce geographical scope, national trends aren't as relevant or realistic.

> Because the percentage of black GitHubbers in management positions is far below e.g. the percentage of e.g. black people with CS degrees in the US

Where's the data for this? Often people talk about these things as assumptions, but I've never seen credible data. Can you share a source?

Furthermore, even the comparison that you mentioned is not accurate. It should start from a funnel like: "the number of qualified black applicants to GitHub with CS degrees", or something along those lines. GitHub can't hire people that don't apply.

Starting from only the people who apply gives a free pass to biased recruitment.

What, you want them to go kidnap black people and force them to apply?

So you don't think it's a problem if their ads say "whites only"? Or if they only advertise to white people? Or are you thinking "nobody would say whites only" and just totally unaware of the many ways that language and targeting can affect who applies to a job, sometimes skewing the pool of applicants in unintended ways? A nice, nobody-meant-it way to bias your candidate pool: only hire people who know an existing employee. White Americans mostly know white Americans.

You're asking for a citation for zero being less than non-zero? :)

He's asking for a citation that the probability of seeing zero is small (p<0.05 if you want to be frequentist about it), contingent on a proper funnel analysis and no bias.

I kinda have the same thinking. Saying that "we have no Black/African-American (...)" seems way more racist that don't saying anything at all. People should get hired according their skills, not race/gender/etc. If a company is focused in hiring non-white males, aren't white males in disadvantage on recruitment processes?

Agreed. These types of reports make it seem like Github cares only about skin color which to me is racist.

This viewpoint genuinely baffles me. You can only manage what you measure, and the process of monitoring a metric isn't the same thing as caring about the metric. I monitor free disk space on all my servers, but it doesn't mean I care primarily about buying more disks. In fact I don't care because I've monitored it and have determined we're okay there. However, as a result of data, if my servers were regularly running out of disk space, then I should start caring.

If these reports determined that there was no evidence of something to be fixed (either the statistics were consistent with the statistics of qualified potential employees, or there were a good reason for them not to be consistent), I would expect GitHub not to care. And yes, if they continued to care about shifting the numbers further past that point, then you could accuse them of caring only about skin color when it's unjustified. But why should a technical company not measure metrics and gather data?

The argument is that making decisions based on race is racism. Github isn't "monitoring" things, they're actively trying to change them, and using race (and other such factors that some people believe we should try to be blind to, or at least agnostic about) as a factor.

Yes. I understand "racism" (at least in this context) to mean "bias in favor of or against candidates based on their race", not "paying attention to race". The latter is something we can hopefully stop doing once we're confident the former isn't happening. I don't understand the reasoning that involves refusing to think about whether the former is happening. (And besides, if GitHub as policy does not pay attention to race, but individual interviews and reviews -- which are notoriously subject to human whim on all sorts of axes -- might be biased on race, you definitely haven't achieved the goal of your system not paying attention to race.)

Again, this seems pretty straightforward for a technical shop... if your goal is, say, "We should not have features on GitHub for Desktop that preference one platform over another," then the way to get to that goal is you add testing processes that explicitly try out things on Windows, Mac, and Linux. Even if you develop your software with some platform-agnostic library (which is a perfectly reasonable way to achieve the goal), you test the software in a non-platform-agnostic way, precisely because you want to know if you got it wrong. You don't say "We're platform-agnostic" and remove the OS field from bug reports, unless you are 200% sure that your code is platform-agnostic even when things go wrong. If you see lots of cross-platform bugs, fine! You still leave it in there so that, one day, if somehow things are disproportionately broken on Windows, you're able to actively respond to that.

I think not talking about race is more damaging than talking about race in many contexts. I also think that a decrease in institutional privilege (white male, in this case) is not discriminatory.

I also don't think "quotas" are a good solution, e.g., "Hire one more Mexican woman and we're good!" But acknowledging some statistical realities and taking a close look at what got you there isn't a bad thing.

Behavior introduces racism, not talking. Talking about it means we might find a solution or a mitigation.

You make it sound like the problem would go away if we'd just stop talking about it. That's not a solution; that's just sweeping it under the rug.

Talking is a behavior, is it not?

The problem ises some people know that they can discriminate against people based on skin color. They learned that certain groups had certain stereotypes. They learned these "facts" because people around them reinforced that these groups were different by talking about how the groups were different.

Talking about race is only going to beget more racism at this point.

"A decrease in institutional privilege"? What are you people smoking?

The choices you make affect real _people_--not institutions; hiring someone over someone else based on some relatively arbitrary superficial trait--thinking that you know that person and what they've lived through--has serious effects on that individual--not an institution. But you choose to make those choices, because you perpetually believe you know things that you do not.

> People should get hired according their skills, not race/gender/etc.

And what do you do when you find out that you're not actually hiring people according to their skills? Like, when that's what you're trying to do—and believe that you're doing—but the data reveals the opposite?

Try to improve their hiring process without focusing on hiring minorities?

It's awkward that all these statistics seem to focus on the output of system without any regard to the input. There's a passing comment about "improve recruiting" but there's no details presented as to who is actually applying.

The hiring process/system is very complicated an difficult to reason about in an objective manner, mostly because it's humans dealing with humans, compounding our human-y traits, non-objectiveness being a hallmark.

I don't see how I'm supposed to take seriously an initiative at attempting to uncover/rectify biases of a system with only the output taken into account. It just seems silly.

This strikes me as centralized planning for jobs. It seems like the best way to go about this is focusing on the pool of applicants and making sure everyone gets a fair shake. Making assumptions about the outcome of those hiring decisions independent of the actual applicants seems neither fair nor good for business.

If your organization's demographics don't match the demographics of the broader population of qualified candidates, there's a good chance it's because you're biased (intentionally or not) in your recruiting or evaluation.

In that sense, "focusing on the pool of applicants and making sure everyone gets a fair shake" are pretty much exactly what diversity efforts are about.

By their wording, it seems like they're more focused on the outcome of the hiring process. No black managers is the outcome, and they said that's unacceptable.

If you look at the process instead of the outcome, it might not be the same. If the scenario was that no qualified black people applied for management positions which resulted in not having any black managers they could still come back to the public and say "We have nothing to apologize for, we tried to hire black managers but none applied" instead of "We failed, we don't have any black managers and it's our fault".

It's good to be unbiased in your hiring practices, but you can only shoulder so much blame. There are many factors that play into the process and you can only accept responsibility for your own actions.

How do you evaluate whether your process is good or bad without looking at the outcome it produces? While I grant that even the best/fairest processes can fail on individual cases, if your aggregate result is poor, what other conclusion can you draw except that your process is broken?

To take your example, if the scenario was that no qualified person from a given group applied, you can look at your process and ask what prevented them from doing so. Maybe there's a perception that your organization is hostile to that group, and you can do something to address that perception. Maybe your job postings are only going up in places that target a specific type. Maybe your 3rd-party recruiters are biased. Whatever the reason is, if the broader population of qualified candidates includes groups that aren't applying to your organization, that's pointing to a process problem you can try to fix (which I very much think of as accepting responsibility for my own actions).

>there's a good chance it's because you're biased (intentionally or not) in your recruiting or evaluation.

I don't think this is necessarily the case. Or even likely the case. For example, we know that the number of blacks, latinos, and women studying CS, etc., is much lower than their proportion in the population. It follows that even the most equitable tech company is unlikely to have demographics which match the population at large. You can't hire black coders in proportion to the general population if black coders form 5% (or whatever it is) of coders while black people form 14% of the general population. The math just doesn't work.

I tried to be careful in my phrasing, which is why I said "broader population of qualified candidates" and not "population at large" or "general population", and "good chance" instead of "certainty", because as you point out there are unfortunately systemic factors outside of your immediate short-term control that will affect the numbers.

But I think a lot of people/companies use this as an excuse to just not even try, and that doesn't make sense. Even if a group's share of qualified candidates is smaller than its share of the general population, that share certainly isn't zero, and once your team reaches a certain size its not hard to check the math and see if your process is producing results in bounds of reasonable estimates. Very frequently it isn't, and that can point to hidden or explicit biases that you can work to remove from your process.

>But I think a lot of people/companies use this as an excuse to just not even try, and that doesn't make sense.

Doesn't it? I doubt that hiring people of different ethnic backgrounds from the same country makes any difference in the performance of a team, so exerting extra effort to do so is likely a net loss for the company.

Let's say I see 2 qualified candidates a week and they accept our offer at a rate of 20%. To ramp up to a team of 12, I'll had to have seen 60 candidates, which will take 30 weeks.

But then I take a look at my process and realize I've got a problem that's keeping 1/3rd of my qualified candidates from even applying. Fix that, and I jump up to seeing 3 qualified candidates per week, which drops my ramp time to just 20 weeks.

The effort I put into fixing my hiring process just gave me 10 bonus weeks of a fully-ramped team. If your company actually knows how to use developers to make money, that could be literally millions of dollars.

It's possible, but you'd have to prove it, and I think that's an uphill battle.

Further, that's not what the previous poster was saying, and not what I was addressing. They said that there's a high degree of likelihood that not reflecting the population at large is the result of bias. That, as I said, is not likely to be true.

> That, as I said, is not likely to be true.

I think it's by far the likeliest explanation given everything we know of human societies and human history (i.e, as a general rule, almost everything is fluid). Moreover, it's the safest: if, as a result of such under-representations, certain groups have less power in society, it's unreasonable not to try to fix it. If people get sick and you don't know if the disease is curable or not, the far safer (and more ethical) assumption is that it is. Both assumptions are not symmetrical.

> If your organization's demographics don't match the demographics of the broader population of qualified candidates

This hasn't been demonstrated.

Creepier than a few centuries of slavery? Creepier than a few centuries of anti-black, anti-woman laws at every level of government? Creepier than nearly universal, measurable, provable anti-black, anti-woman implicit bias today?

Sure guy.

Discriminating by sex and race is what caused those issues in the first place.

lol.