As a Google employee, the reasons I hear is that we'd like the diversity of our workforce to more accurately reflect the diversity of our users, so we can serve them better.
Deleted comment
No to the quotas.
But yes, race in the NFL (and in many sports) is an issue that should be addressed. The old problem of the low proportion of black quarterbacks is slowly being addressed, and the Rooney Rule[1] has had some success with the underrepresentation of black coaches.
Deleted comment
Like most things to do with race there is a tiny, tiny sliver of truth in this, but not enough to explain the outcomes.
sigh
Lets look at sports.
Sprints: The ACTN3 gene[1] is associated with the fast-twitch muscles, and is more prevalent in people of West African descent. Out of the last 7 Olympic 100m mens finals all 56 finalists have been of West African descent. Only 2 out of the top 500 100m race times are by people without African descent.
Case closed? No. The proportion of sprint wins by West Africans is much higher than the prevalence of the ACTN3 gene would explain, and other power sports (notably weight lifting) are dominated by other ethnic groups with a more average distribution of the gene.
Distance running: The Kalenjin tribe (mostly from Kenya) has dominated distance running since the late 1990s. Most of their competition has come from Ethiopian runners and the occasional North African (especially in the women's events)[2].
They do have a genetic advantage: they tend to be very skinny, with long legs and short torsos. But you can find these attributes in any population group. More surprising is that elite non-African runners performance has decreased over time: For example, of the top 10 Marathon times by British athletes only 1 (in 8th position) is since 2000, and 6 are pre 1990 (including the 2 fastest times).
This indicates something else is going on, that can't just be explained by genetics.
[1] http://en.wikipedia.org/wiki/ACTN3
[2] http://www.iaaf.org/records/toplists/road-running/marathon/o... - the best place person without a African heritige is Ronaldo Da Costa at 57.
> Sprints: The ACTN3 gene[1] is associated with the fast-twitch muscles, and is more prevalent in people of West African descent. Out of the last 7 Olympic 100m mens finals all 56 finalists have been of West African descent. Only 2 out of the top 500 100m race times are by people without African descent.
> Case closed? No. The proportion of sprint wins by West Africans is much higher than the prevalence of the ACTN3 gene would explain, and other power sports (notably weight lifting) are dominated by other ethnic groups with a more average distribution of the gene.
It's hard not to conclude that you're purposefully trying to fool someone here. It never occurred to you that more than one gene might be involved in sprinting?
Sure, we know there are lots of things we don't know yet. But we do know generally in almost every case every ethnic group has individuals whose genetic makeup overlaps with those from other ethnic groups.
It's hard not to conclude that you're purposefully trying to fool someone here. It never occurred to you that more than one gene might be involved in sprinting?
My point is that there is a lot more involved in sprinting (and high-performance generally) than just genetics.
One of the best examples of this is the cross country skier Eero Mäntyranta[1]. His family carries a genetic mutation called Polycythemia[2] which means they produce a very high proportion of red blood cells (basically like the EPO doping Lance Armstrong did, but natural).
Juvonen described lab-testing samples of Eero’s bone marrow cells, which produce red blood cells. The plan was to add EPO to the cells and track the creation of red blood cells. A typical human’s bone marrow cells won’t start making blood cells in the lab until EPO is added. But Eero’s bone marrow cells began the process of creating red blood cells before Juvonen could even stimulate them with EPO. Whatever tiny speck of EPO that was already in the sample was enough to keep the red cell factories humming. Eero’s body was hypersensitive to even trifling traces of EPO.[3]
He was a very successful cross country skier (Olympics: 3 Gold, 2 Silver & 2 Bronze over 3 games), but he was still beatable, even with this unprecedented genetic advantage.
A quote from [3]:
This is why any sensible person, in response to the questions: “Is there a genetic test for performance? Is there a performance gene?”, can only ever answer “No, it’s too complex to reduce to a single gene, and tests cannot assess this level of complexity. Yet”.
Mäntyranta is about as close as it gets to that. However, despite that, I want to point out that even with this mutation, one which saw his hemoglobin levels increase by around 50% compared to the typical male, he still didn’t wipe the floor with everyone, every time. Yes, he dominated his events in unmatched fashion in 1964, but he still got beaten, and that single mutation did not confer invincibility, only advantage. That’s again an illustration that performance is multi-factorial, and it suggests that many other factors are responsible.
[1] http://en.wikipedia.org/wiki/Eero_M%C3%A4ntyranta
[2] http://en.wikipedia.org/wiki/Familial_erythrocytosis
[3] http://www.sportsscientists.com/2013/12/eero-mantyranta-finl...
(1) Not enough black quarterbacks
(2) Not enough black coaches
and those are issues that need to be addressed?
How is the low proportion of black quarterbacks a bigger problem than the ultra-low proportion of white non-quarterbacks?
It would be if there was any other indication of discrimination against whites than the numbers. In the case of coaches and quarterbacks there was.
When I wrote that I wasn't aware of any evidence showing discriminatory hiring practices affecting white players, but I have since found an allegation of one incident[1].
Given that, I'm changing my position[2]: Yes, the NFL should act on this. I'm not sure what an appropriate course of action is (given that quotas don't work well), though.
[1] http://rollingout.com/sports/stephen-a-smith-discusses-rever...
[2] Strong opinions, weakly held...
I agree with you that I don't think the NFL is discriminating against white players. They just want the best, period.
What I have issues with is when a company's racial makeup shakes out to be majority white (funny how that can happen when you draw from a pool that's majority white) and the first instinct is to say "this _must_ be because of bias."
a) White males are better at software engineering and that's why the majority of software engineers are white males (there is no scientific evidence to support this whatsoever).
b) There are socio-economic factors that lead more white males into software engineering than anyone else (there is ample scientific evidence to support this theory).
Clearly there is a bias at work somewhere. It's not necessarily a hiring problem, but rather one that starts further up the "supply chain" (if you will). We're trying to fix that.
Professional football is by far the most respected profession here. Top of the food chain. Of course, most teenagers recognize they aren't going to make the high school team, let alone make it into the professional realm. If you do make the team, well, try as hard as you can. And of course, most people still don't make it, there's just not enough demand for pro football players. But it's still a huge thing that people aspire to. If you do make it, everyone is going to speak of you with awe. There's a couple guys who grew up near where I did and every single person who knew them or knew of them (anyone you meet will eventually bring them up in conversation) speaks of them with total reverence - they _made it_.
Unless your parents are wealthy and/or well-educated, most people don't have a "career" as such in mind - just jobs. Football stands out - especially since the perception is that the only other professional careers boil down to things like "doctor" or "lawyer" which (so the thinking goes) require highly exceptional intelligence beyond the ken of most people, and anyway, they aren't nearly as glorious or high-paying. So it's all about aspiration - and in some places, pro sports is at the top, even if you or I think it constitutes a terrible or unrealistic career choice.
Now, in my experience, that aspiration is strongly correlated not so much with race but with your family's educational history, professions, and income. "Rich" kids (parents with income > 60k) grow up in an environment where they really understand there are other opportunities out there. They're more likely not just to go to college, but grow up believing that college is an inevitability and part of life. A ton of kids never even get that far. I think that's the kind of kid he's talking about.
Deleted comment
The biggest diversity win for them would be expanded hiring in all groups, but greatest in those currently underrepresented.
OK, so it sounds like they've got a fix that'll only take a couple decades to implement.
In the meantime, how do you think Google is going to improve their minority percentages in the absence of qualified minorities?
Google is not going to in any way limit the intake of qualified candidates. It is mission critical for us to hire as many talented people as possible. Most teams at the company really need more people.
But it's nice of you to assume the worst of us in that hypothetical situation.
I would assume that when that happens we would raise the hiring bar. And by that time there would be more diverse candidates in the pipeline, so we would naturally get a more diverse spread of new hires.
I strongly doubt we would sacrifice our long held policy of "hiring the best" for the sake of diversity. It flies in the face of our corporate culture.
My understanding of Google's diversity efforts is that they are centered around a) encouraging candidates from more diverse backgrounds to apply to Google (many, particularly women, do not believe they are as good as their male counterparts for social reasons), and b) encouraging people from diverse backgrounds to get into computer science. I have been involved in both these efforts at Google. And, to be quite honest, only a sociopath would see a problem with these initiatives.
Yeah, but this is about Google, at the moment. Not about "typical companies" or "Google in some speculative future where they are money or task constrained, rather than employee-pool constrained".
No, I don't. The discussion is about Google's current state, and Google current plans to deal with diversity given that current state. A change in circumstances (and Google going from "hire every qualified person we can get our hands on" to "stop hiring anyone" would be a radical change in circumstances) would be an occasion for a different discussion.
> Are you really that naive to think Google will continue to expand indefinitely?
I think that Google's concerns about and options for approaching diversity would be different if Google was funding or task constrained rather than employee-pool- constrained, and that the proper focus of discussion is the real world that exists, not some speculative future that may or may not exist in the lifetime of anyone discussing it.
> They absolutely will hit a point when they can't just leave the number of new hires uncapped.
Its almost certain that sometime before the end of human civilization that that will happen for Google, but the time to discuss whether diversity should remain a concern and what steps are appropriate to address that concern in a funding-constrained firm will be appropriate with respect to Google when that occurs. However, it isn't the case now, so its just a distraction from the issue today.
Especially since I'm sure that -- as much as Google would hope to avoid that circumstance as long as possible -- Google would also prefer to acheive an acceptable demographic profile in its workforce before it becomes constrained in that way (and, insofar as there are business as well as humanitarian reasons for its diversity concerns, sees acheiving a good diversity profile as part of a strategy to continue growing in a way that does not result in them becoming resource- rather than employee-constrained.)
So? Some things takes lots of time to do right.
> In the meantime, how do you think Google is going to improve their minority percentages in the absence of qualified minorities?
I don't think Google has said anything which indicates that they plan to "improve their minority percentages in the absence of qualified minorities."
X is an over-represented group of Googlers threatening their diversity. Let's say it's males. Y is, let's say, women. Z is the total number of current Googlers.
In this case, 0.7Z = X and 0.3Z = Y. (70% men, 30% women).
Still with me?
If Google were to continue hiring X at the current rate (i.e., 70% of all candidates hired), then the diversity of Z will not change with respect to X and Y.
Therefore, in order to achieve the desired balance, Y must be >> X, be that through hiring (i.e., hire more Y than X) or firing (fire more X than Y). Rinse and repeat for other over- and under-represented groups at Google. I'm not sure why this concept is lost on you.
If Google has 100 X's and 10 Y's and hires 10 X's and 1 Y every year, the ratio of X's to Y's stays the same. If Google increases the number of Y's they hire each year to 2, the ratio of X's to Y's decreases. Not sure what's so hard to comprehend about this.
According to Google's numbers, whites are underrepresented at Google compared to the US at large.
What? What did I say?
Yes, european whites make up roughly 72% of the US, http://en.wikipedia.org/wiki/Demographics_of_the_United_Stat..., and 61% of Googlers.
If anything, you should be wanting to see less Asian faces overall (insanely, dramatically overrepresented, 30% of Googlers compared to 4.7% of Americans), and want to see more black faces and more white female faces.
Like you, I enjoy engaging with a lot of different sorts of people. I was mostly just commenting that I've found a lot more success doing so outside the workplace in my own personal life. Maybe it's the places I've worked? Hard to say.