From the original memo:
> However, to achieve a more equal gender and race representation, Google has created several discriminatory practices:
> ...
> Hiring practices which can effectively lower the bar for “diversity” candidates by decreasing the false negative rate
No he hasn't. Rather, he's implying that 'non-diverse' hires have a higher bar to to pass, and that by enforcing diversity quotas you tend to end up hiring exceptionally qualified people from the 'non-diverse' group, which can have the opposite of your desired effect, and perpetuate stereotypes about the performances of people from different groups.
For example, let's say you have 100 people interviewing for 10 positions. Of that, due to current representation of women in tech let's say the candidates comprise 20 women and 80 men.
Let's also assume that 25% of those interviewing meet the technical standards a.k.a 'the bar' required by Google (that makes for 5 women and 20 men).
Google however is only hiring 10 positions, and in order to meet diversity requirements will hire 5 women and 5 men (50/50 split).
Each one of the people hired is qualified to work at Google and meets the bar (none of them are underperforming).
Without any diversity requirements a candidate would need to be in the top 10% of other candidates to get a position however now, for the women, they only need to be in the top 25% of other woman candidates to get the position, whereas the men need to be in the top 6.25% of male candidates.
The bar has been effectively lowered for the women candidates and raised for the male candidates, despite all of them being worthy and capable of working at Google.
If you are always hiring equal numbers of people from both groups, but from one group you are always hiring the top 6.25% of candidates, and the other group you are hiring from the top 25% of candidates, then over time you'll start to find that on average, one group has a larger percentage of high performers than the other.
A professor of social psychology at Rutgers said:
> The author of the Google essay on issues related to diversity gets nearly all of the science and its implications exactly right.
An evolutionary psychology professor said:
> For what it’s worth, I think that almost all of the Google memo’s empirical claims are scientifically accurate. Moreover, they are stated quite carefully and dispassionately. Its key claims about sex differences are especially well-supported by large volumes of research across species, cultures, and history. I know a little about sex differences research. On the topic of evolution and human sexuality, I’ve taught for 28 years, written 4 books and over 100 academic publications, given 190 talks, reviewed papers for over 50 journals, and mentored 11 Ph.D. students. Whoever the memo’s author is, he has obviously read a fair amount about these topics. Graded fairly, his memo would get at least an A- in any masters’ level psychology course. It is consistent with the scientific state of the art on sex differences. (Blank slate gender feminism is advocacy rather than science: no gender feminist I’ve met has ever been able to give a coherent answer to the question ‘What empirical findings would convince you that psychological sex differences evolved?’)
> But it is not clear to me how such sex differences are relevant to the Google workplace. And even if sex differences in negative emotionality were relevant to occupational performance (e.g., not being able to handle stressful assignments), the size of these negative emotion sex differences is not very large (typically, ranging between “small” to “moderate” in statistical effect size terminology; accounting for less than 10% of the variance). So, using someone’s biological sex to essentialize an entire group of people’s personality would be like operating with an axe. Not precise enough to do much good, probably will cause a lot of harm.
> Again, though, most of these sex differences are moderate in size and in my view are unlikely to be all that relevant to the Google workplace (accounting for, perhaps, a few percentage points of the variability between men’s and women’s performances). Sex differences in occupational interests, personal values, and certain cognitive abilities are a bit larger in size (see here), but most psychological sex differences are only small to moderate in size, and rather than grouping men and women into dichotomous groups, I think sex and sex differences are best thought of scientifically as multidimensional dials, anyway
> Within the field of neuroscience, sex differences between women and men—when it comes to brain structure and function and associated differences in personality and occupational preferences—are understood to be true, because the evidence for them (thousands of studies) is strong. This is not information that’s considered controversial or up for debate; if you tried to argue otherwise, or for purely social influences, you’d be laughed at.
> Sex researchers recognize that these differences are not inherently supportive of sexism or stratifying opportunities based on sex.
That the empirical claims are accurate doesn't mean that they are relevant to the (memo) author's motive for writing it, whatever that is.
*Citation needed.
If you're suggesting they were subjected to different practices, contrary to what HR at Google is stating, then you better cough up evidence. Do the women and minorities at Google perform worse than White males?
This part in particular is worth considering:
"..if minds differ across sexes and races enough to justify diversity as an instrumental business goal, then they must differ enough in some specific skills, interests, and motivations that hiring and promotion will sometimes produce unequal outcomes in some company roles. In other words, if demographic diversity yields any competitive advantages due to psychological differences between groups, then demographic equality of outcome cannot be achieved in all jobs and all levels within a company."
Create STEM schools and give kids computers in elementary school. Use more fun robotics in the classroom. Teach programming literacy. Let kids make games.
Of course the caveat to these approaches is that it takes ~13-17 years from kindergarten until high school graduation / college graduation to when the candidate can then use the skills by starting a tech job. That said, I think upstream educational opportunities like these are the best long-term solution to diversifying the tech population.
1. Making the workplace less hostile to that group. The code of conduct and this firing both seem to addressing this point.
2. Making sure salaries are equal for equal work. Google might have some problems in this area.
3. Investing in the hiring pipeline (education) to make sure bias isn't occurring earlier in the pipeline.
Those are just a few ideas. Someone really committed to that goal could surely find more ideas.
Then in a decade you'll have a much wider pool of talent to select from and can just take the best while appeasing diversity advocates.
The same could be done for racial minorities.