Slashdot Mirror
Google Diversity Report Straight Out of 'How To Lie With Statistics' Playbook
theodp writes: Among the books recommended by Bill Gates for beach reading this summer is How to Lie With Statistics, the published-in-1954-but-timely-as-ever introduction to the (mis)use of statistics. So, how can one lie with statistics? "Sometimes it is percentages that are given and raw figures that are missing," explains the book, "and this can be deceptive too." So, does this explain Google's just-released Diversity Report and the accompanying chock-full-o-percentages narrative (find-all-%-image), which boasts "the Black community in grew [sic] by 38 percent", while the less-impressive raw figures — e.g., the number of Google employees increased by 5,928, but the ranks of Black females only increased by 35 (less than 0.6% of the net increase) — are relegated to a PDF of its EEO-1 Report that's linked to in the fine-print footnotes? To be fair to Google, Microsoft, Twitter, Apple and Amazon didn't want people to see their EEO-1 numbers, either.
The big lie, I guess.
If you imagine that you are tired of hearing about it as a reader or tech employee, just imagine how it might be for the people whose job it is to make this bettter at the supposedly forward thinking tech giants.
How are those numbers coming, Jim?
Well, we've hired as many somewhat qualified people as we can find, and it's still not enough. Can we count the cafeteria employees again this year?
Happiness in intelligent people is the rarest thing I know.
Ernest Hemingway
My father told me that when I took math classes in college, that Statistics I will teach me everything I really needed to know about the subject, but that Statistics II would teach me how to lie with what I learned. He was not incorrect. There's so many ways to manipulate the data that I find it very, very difficult to trust ANY stats that I find in the news without also having access to the raw data, the methodology, questions used, selection process, etc., etc., etc.
Love sees no species.
Google hires people based on talent. Women and minorities are under-represented in the technical and engineering community. That is a fact of life. Until more women and minorities CHOOSE to enter this field, getting a "diverse workforce" would have to mean you exclude more qualified white males in order to hire less qualified minorities and women.
Think about that for a moment. Suppose hospitals did things this way? If you need critical brain or heart surgery, do you want your surgeon to be one of the best in his or her field, or one that was a "diversity hire"?
Until you're comfortable with the second option, this "diversity" idiocy needs to stop. It's one thing to exclude perfectly qualified candidates because they're female or minority. It's another thing to make that the primary reason you're hiring them instead of making sure they're the best qualified for the job.
In the end they will lay their freedom at our feet and say to us, Make us your slaves, but feed us. - Fyodor Dostoyevsky
What evil is Google perpetrating, exactly? Trying to present irrelevant data in a positive light to people who pretend it is relevant?
appended to the end of comments you post, 120 chars
Related to that, however, is the question of what hormonal influences may arise. For one example (of many possible), with males, you often see more aggression, and (obviously) with females, less. Pretending there can be no relevant differences WRT job performance is not an optimum approach. Furthermore, interactions between the people of significantly different sexual identity are of inherently different natures. Much as the incoherent would like you not to believe it, the vast majority of us are sexual creatures. We are naturally and unavoidably affected by other concerns than the specifics of today's TPS report.
Same thing goes for age, various cultural influences, parent or not, single or not, personal maintainance, presentation, health, mobility, superstition, depth of education, and means of education (conventional, autodidact, on-the-job, etc.)
Because of these truths, consideration should be given to such factors. And of course it is, and always will be. But mostly because of the law, much of this is now sub-rosa, which is entirely a bad thing -- a bad thing that at least partially offsets the benefits of the law overriding (or at least attempting to override) people who operate using a chain of reasoning that primarily incorporates blind prejudice rather than "how will this affect job performance?"
Politically correct often means "poorly thought out and mostly harmful." When there are differences, there are differences. Pretending otherwise doesn't make such things go away. It just makes them harder to deal with.
I've fallen off your lawn, and I can't get up.
Comparing "murder" with "reporting the race of Google employees in a way you don't like" is a little hysterical, don't you think?
"Google is being evil here"
Uh, Yah... they are being real evil.
They should fire a bunch of white people and hire a bunch of non-white people based solely on the color of their skin.
That would make them not evil.
You are under the impression that people tell the truth. Very rarely do people tell the truth. The whole Political Correctness gig is simply a "how to avoid telling the truth". We are lied to daily, by almost everyone that has media outreach, from broadcasters to politicians to advertisements to government officials to business leaders.
The fact that most Americans still "trust" anyone at this point is a testament that lying still works. AND If you say it convincingly enough (wagging finger ... "I did not have sexual relations, with that woman, Ms Lewinsky") it actually works!
The best liars are the ones that you thank after they've lied to you. "Thank you sir, may I have another?"
Agent K: A *person* is smart. People are dumb, stupid, panicky animals, and you know it.
To be demographically representative of the US population, they would have to fire lots of Asians, and hire more Caucasians and African Americans, since the latter two groups are both statistically underrepresented.
In this case (there are certainly others), they are using deceptive reporting to mislead people on the current state of affairs. Ask yourself why they would do this. The answer isn't "because they are angels."
Probably because they, like many other tech companies, are getting incessantly railed on an issue that is out of their control.
I don't know about you or any of these other SJWs, but I went to college for an IT career, and I only recall seeing at most one or two black people to a class the size of about 30 in any of my technology classes. In other classes I took (mainly the general requirement classes) there were more. (Most of the ones I met were either going for legal or service industry management careers.)
For whatever reason, most of them don't care to pursue a career related to technology. That isn't Google, Microsoft, Amazon, or anybody else's fault. Meanwhile they have to catch shit about it all the time, and pay ransom money to Al Sharpton (who himself is the real lying sack of shit.)
The same can be said of women, by the way. As another anecdote, two of my cousins are currently wanting to get IT jobs, but their sister wants to become a dentist, and that isn't due to any different treatment by their parents (they buy her as much computer stuff as they buy for her brothers. In fact she often asks for and receives more expensive Apple phones/tablets where they get Android devices.)
I clicked the link hoping for actual evidence, but I got more speculation.
To answer the questions posed at the post linked, what's happening is that gender lunatics along with other special interest groups are simply hammering tech right now with the equivalent of "Have you stopped beating your wife yet?"
As far as racial issues, hoo boy. Yes, there are quite real problems here, but the problems are at the K-12 stage, not at the workplace. If these groups were serious about changing things, they'd be out in communities trying to help families realize value for education and to get schools in impoverished neighborhoods more funding and better resources.
I am far more receptive to the overall race issue. We can't expect centuries of slavery and Jim Crow laws to just magically sort itself out. BUT: tech companies and workers can do nothing about it, short of donating to some initiative that fits what I wrote above. If there is no evidence that these companies are discriminating against actual applicants, what the fuck are you asking them to do?
What I need to change my mind are some actual demographic data. How many black applicants with this GPA from X college applied and were hired vs. how many white or Asian applicants with the same GPA from same X college applied and were hired? That's how you generate reasonable suspicion of racism. Without knowing how many black applicants who at least on paper appear to be equal to their white and Asian colleagues, even knowing the number of employed blacks is meaningless to me.
Then there's the gender lunatics.
My first question, although it would amount to a rounding error, concerns trans women. How are trans women reported? Do we go by lived gender (i.e. name's changed, on HRT)? Do we go by bottom surgery status? Do we go by birth certificate gender, which, depending on the state of birth, cannot be changed even after bottom surgery? Or do we go with the feminist view as stated by Janice Raymond and the Michigan Womyn's Music Festival (i.e. trans women are not only rapists because they're men, but also metaphysical rapists because they have the audacity to have the same body parts as women)?
So, ok, rounding error aside, but I'd like to segue into the real point that the preponderance of trans women who are in tech raises the question of why are there very few cisgendered women who go into tech? Sure, I'm being a hypocrite here because I don't have the demographic data to back up my assertion, but visit any gallery of successful trans women. Sysadmins! Programmers! EEs! DBAs! More sysadmins and programmers!
There is no economic disparity in childhood between daughters and sons, unlike the race issue.
Do you have a good theory we can use to explain this?
Gender lunatics across the board from feminists to MRAs all take a shit on trans women. Yet, you'll find trans women in tech. Why? What is the difference here between cisgendered and transgendered women?
-- kurenai.tsubasa (not bothering to register this account here because fuck beta)
Its relevant to knowing if their hiring practices are biased for or against people with a certain skin colour, for example...
No, it's not. To do that, you'd need an analysis that includes much more than just who is working there now. Two huge variables you're leaving out:
1) How many minorities actually applied for a job there?
2) Of those, how many were actually qualified?
Go take your zero knowledge SJW outrage to twitter and/or wordpress where it belongs.
The first two people I hired to work at my business were females. The first was a stenographer/secretary. The second was a programmer of some note that I stole from a buddy (he did not mind too much) who owned his own company as well. The third was a black male who was also a programmer. Then came a bunch of females as a sales force - being held accountable to a female. For quite some time my other business friends called my shop, "The Bunny Ranch."
Actually, the first person I hired was Amerindian, Black, and White. It was me. I had an asshole for a boss too. He made me work long hours, sometimes without pay, and seldom paid me for all the hours I put in until long after the company was secure.
Anyhow, my point has nothing to do with me wanting diversity. In fact, that was the furthest thing from my mind. I wanted to hire the best, pay them well, and encourage low turnover by incentivising (spell check says that is not a word, if it is not then I declare it as such now) remaining with the company. I was easily approachable (then HR was) and if one felt they were inclined to leave we would do our best to ensure that they were given reasons to remain. We paid more by default, we offered better benefits by default and immediately upon hire, and we offered "telecommuting" long before such was a word. Due to the expense we paid for their connection and hardware as well - allowing them to keep the hardware (even if they resigned) and generally left the connection running for some time as well. This was seldom applicable though, and this was a good thing.
It is not diversity that matters (though it can happen by accident as you see above). What matters is merit, low-turnover, and an employer that does not treat you like an asset or a piece of meat.
"So long and thanks for all the fish."
In this case (there are certainly others), they are using deceptive reporting to mislead people on the current state of affairs. Ask yourself why they would do this. The answer isn't "because they are angels."
Nope. It's because it was a USA Today hit piece from 6 months ago, and educators, parents, and guidance counselors don't want to take responsibility for the input to the pipeline, and it's a slow news day.
They don't release the numbers because they don't want to be blamed for them, when they can only take whatever output comes out of the pipeline.
It's not like there are huge numbers of PhD CS people in the underrepresented minorities just sitting around twiddling their thumbs: everyone knows that the only people who get discriminated against for these jobs are people who are old (sorry... "not 'digital natives'... must use the code words), or who learned by doing, and so have no reasonable credentials with which to protect their employers in the event of a lawsuit.
It's also not like the CollegeBoard would not *gleefully* take the money of anyone who wanted to pay for the AP Computer Science test, or any other freaking AP test, period: they will happily take *all* your money if you are willing to give it to them.
No the problem is the input to the pipeline, and it's not being addressed, and so on slow news days, you get attack pieces on the people on the other end, as if they could magically make an Comparative Literate graduate into a software engineer.