Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Brief thoughts on the “Google memo”

juliagalef.com

Post reply on HN
Print view View original

41–50 of 194 posts

Re: Brief thoughts on the “Google memo”

#41

Post by kobeya »

jamesrcole wrote: post #35
my_first_acct wrote: post #4

Recycling (and slightly modifying) a previous comment [0]: Certain topics, that otherwise might be interesting to discuss, are surrounded by minefields. One such topic is the distribution of intellectual ability within subgroups of the population. This minefield was not put in place by a repressive government. Nor was it secretly put in place overnight by a fanatic band of social-justice zealots. My observation, whic…

> I will invoke the principle of Chesterton's fence [1]: Before you talk of removing the mines, you need to show that you understand why the minefield was created in the first place, and you need to explain why now is the time to remove it. That's a bit different to what the Wikipedia page describes it as. It describes it as being to important to understand the reasons for X before trying to change X. It doesn't say…

> That's a bit different to what the Wikipedia page describes it as. It describes it as being to important to understand the reasons for X before trying to change X. It doesn't say anything about needing to show that you understand it, and needing to explain why now is the time to remove it.

That seems like the same thing to me...

Re: Brief thoughts on the “Google memo”

#42

Post by kobeya »

nippples wrote: post #38
DataWorker wrote: post #36

What happened to the good old days of vim vs emacs? The real basilisk is believing basilisks exist. Heterodoxy strengthens us all and if you chose to exclude the smart people because their thoughts are unhelpful, or patently offensive, you might find yourself without any allies when you need them most.

> What happened to the good old days of vim vs emacs? Whoa, whoa. Are you trying to start WW3?

Nah, it's just a precursor skirmish for spaces-vs-tabs.

Re: Brief thoughts on the “Google memo”

#43

Post by imron »

shadowmint wrote: post #11

> His hypothesis implies that personality differences mean a smaller percentage of women will be interested in and/or qualified for a Google job than men, but that doesn’t mean any of the women at Google fall below the “qualified” threshold. I think its fair to say that not everyone shares that view. This is probably the key to controversy in my mind. You could read it that way. However, you can choose to interpret t…

> 'Hiring practices which can effectively lower the bar for “diversity” candidates by decreasing the false negative rate'

Imagine you have 100 people applying for 10 positions, and the demographic breakdown of applicants (based on current trends in tech) is 20 women and 80 men.

Let's also say that being male or female doesn't give you any special aptitude, and that the aptitude of all candidates regardless of gender follows a normal distribution along a bell-curve.

Because of that and because of Google's standards, let's say that only 25% of all candidates have the skills and qualifications necessary to work at Google, which breaks down to 5 women and 20 men.

Once again assuming a normal distribution with neither men nor women over or underrepresented in skill level, then without any sort of correcting for diversity, candidates would need to be in the top 10% of all candidates to receive a job offer (10 / 100). , this is 2 women and 8 men, leaving us with a false negative rate (i.e. people skilled enough to work at Google but who get declined) of 3 women and 12 men.

But with an eye for diversity, Google decides to hire all the qualified women (5), with the remaining positions taken by men, so 5 and 5, a 50/50 split.

Each of the candidates is qualified and capable of working for Google, but the false negative rate for women has been reduced from 3 to 0. This is what he meant by an effectively lowered bar by decreasing the false negative rate. The women now only need to be in the top 25% of all women candidates, whereas the men need to be in the top 6.25% of male candidates.

There's nothing false or wrong about that statement, it's just maths that happens as a result of the diverse hiring practices skewing the distribution of candidates, with the bar being 'effectively' lowered for the diverse candidates from top 10% to top 25%.

This has the unfortunate side effect that over time, if you are taking the top 6.25% from the non-diverse group (men in this case), but the top 25% from the other group then the non-diverse group will eventually have on average more high-performers (which affects things like salaries, promotions and more).

It's not to say that men are better, just that the hiring practices have been structured to select men from a higher performing percentage of the population. You would get the same results in favor of women if you were only selecting the top 6.25% of women vs the top 25% of men.

Re: Brief thoughts on the “Google memo”

#44

Post by my_first_acct »

jamesrcole wrote: post #35
my_first_acct wrote: post #4

Recycling (and slightly modifying) a previous comment [0]: Certain topics, that otherwise might be interesting to discuss, are surrounded by minefields. One such topic is the distribution of intellectual ability within subgroups of the population. This minefield was not put in place by a repressive government. Nor was it secretly put in place overnight by a fanatic band of social-justice zealots. My observation, whic…

> I will invoke the principle of Chesterton's fence [1]: Before you talk of removing the mines, you need to show that you understand why the minefield was created in the first place, and you need to explain why now is the time to remove it. That's a bit different to what the Wikipedia page describes it as. It describes it as being to important to understand the reasons for X before trying to change X. It doesn't say…

You're correct that I went a bit further than Chesterton himself did. But not a lot further.

By the way, I linked to the wrong Wikipedia entry. The one I linked to (and that you found the strange quote in) appears to be an internal document, explaining the principle of Chesterton's Fence to would-be Wikipedia editors. I should have linked to the Wikipedia article instead: https://en.wikipedia.org/wiki/G._K._Chesterton#Chesterton.27...

Re: Brief thoughts on the “Google memo”

#45

Post by kobeya »

shadowmint wrote: post #11

> His hypothesis implies that personality differences mean a smaller percentage of women will be interested in and/or qualified for a Google job than men, but that doesn’t mean any of the women at Google fall below the “qualified” threshold. I think its fair to say that not everyone shares that view. This is probably the key to controversy in my mind. You could read it that way. However, you can choose to interpret t…

He specifically said 'decreasing the false negative rate', not 'increasing the false positive rate.' That is explicitly worded (he could have been more succinct) to imply that all people hired by Google are good enough to work at Google. He is saying that IF you are already good enough to work at Google, then your odds if you are from a protected minority group are better than if you are not.

You are right that people interpreted that sentence differently, but those interpretations were wrong.

Re: Brief thoughts on the “Google memo”

#46

Post by siliconc0w »

my_first_acct wrote: post #4

Recycling (and slightly modifying) a previous comment [0]: Certain topics, that otherwise might be interesting to discuss, are surrounded by minefields. One such topic is the distribution of intellectual ability within subgroups of the population. This minefield was not put in place by a repressive government. Nor was it secretly put in place overnight by a fanatic band of social-justice zealots. My observation, whic…

Sure it's uncomfortable to talk about but it's becoming increasingly important that we do. Without understanding why a phenomena like the STEM/leadership gap exists we can't hope to fix it. Diversity programs really aren't working - Google is at 80/20 for tech which is just about the same as the industry (78/22) (Department Of Labor). My policy is 'Fuck Chesterton Fences'(FCF). I don't want to live in a world of rusting fences.

Re: Brief thoughts on the “Google memo”

#47

Post by throwaway91111 »

maehwasu wrote: post #12

> However, his overall claim that there exist personality differences between genders that differentially affect men’s and women’s interest in and aptitude for tech jobs, is what people are mostly getting mad at. And that’s a claim that seems plausibly true. Not obviously true, but also not a claim you would be justified in emphatically dismissing as false, as many people have, including Google, who called them “inco…

Where are you getting that 20% from?

Re: Brief thoughts on the “Google memo”

#48

Post by jamesrcole »

kobeya wrote: post #41
jamesrcole wrote: post #35

Earlier quoted context omitted.

> I will invoke the principle of Chesterton's fence [1]: Before you talk of removing the mines, you need to show that you understand why the minefield was created in the first place, and you need to explain why now is the time to remove it. That's a bit different to what the Wikipedia page describes it as. It describes it as being to important to understand the reasons for X before trying to change X. It doesn't say…

> That's a bit different to what the Wikipedia page describes it as. It describes it as being to important to understand the reasons for X before trying to change X. It doesn't say anything about needing to show that you understand it, and needing to explain why now is the time to remove it. That seems like the same thing to me...

Understanding X is one thing, having to provide a demonstration of that to others, that is considered adequate to others, is another.

Understanding the reasons for X before trying to change X is separate from needing to show why the present time is right for changing X.

Re: Brief thoughts on the “Google memo”

#49

Post by mercer »

matt4077 wrote: post #28

Yes, there may or may not be biological differences. Or cultural differences. In fact, it's almost certainly both, and it really doesn't matter. The single most important argument people like this ex-Googler never seem to understand is that they're confusing what is with what ought to be. If there's one defining characteristic of humanity, it's our ability to overcome biology (the 'hardware',so to speak), using our u…

I'm not sure the ex-Googler would disagree with you on this, though. There's even a section titled 'Non-discriminatory ways to reduce the gender gap' that clearly indicates he at least is not opposed to reducing the gender gap.

Re: Brief thoughts on the “Google memo”

#50

Post by throwaway91111 »

matt4077 wrote: post #28

Yes, there may or may not be biological differences. Or cultural differences. In fact, it's almost certainly both, and it really doesn't matter. The single most important argument people like this ex-Googler never seem to understand is that they're confusing what is with what ought to be. If there's one defining characteristic of humanity, it's our ability to overcome biology (the 'hardware',so to speak), using our u…

Part of the problem is that even traits that are linked to better performance should not be considered qualifications.
Post reply on HN