Postscript on Google and James Damore

I got lots of response, both public and private, to my blog post titled “Google’s Bad Boy,” on Google’s firing of James Damore, who had written a negative analysis of Google’s diversity hiring policies, especially regarding women. I must say that those responses, like the coverage of incident (pro and con) in the media, exhibited a lot more emotion that careful, impartial, open-minded thinking. This clearly is a topic on which different, reasonably-thinking people can come to different conclusions, but unfortunately emotion has tended to prevail in this instance.

A number of readers were upset that I had defended Google’s firing Damore. The fact is, though, that I didn’t say any such thing. On the contrary, when the San Jose Mercury News interviewed me last Monday (at a time that it was not yet known whether Damore would be fired), I stated, no, Damore should NOT be fired. (The reporter ultimately did not use my remarks.) By the time I wrote my post on Wednesday, Damore had been fired, and it had become clear that Google had been forced to fire him for legal reasons, whether they wanted to or not. So I did not comment in my blog on the propriety of the firing, just not an interesting issue to me at that point.

Though missed by many readers, my prime objection to Damore’s essay concerned the blind faith he placed in biological scientific research that supports the Nature side in “Nature vs. Nurture.” Hardly surprising, given his own background in biology research, but in fact that means that he should be extra vigilant about confirmation bias, which he was not.

Contrary to Damore’s unquestioning acceptance of science, there is fairly wide agreement today that science is in a crisis. Plug the term “reproducibility crisis” into (pardon me) Google, and you will see a ton of statements from top researchers and editors of scientific journals who are really bothered by the sorry state of science. The fact is that people have started to realize that Emperor Science has no clothes. There is lack of transparency, sloppy handling of data, erroneous use of statistical methodology, failure to check for inter-laboratory variation, unconscious assumptions, and last but not least, out and out fraud, which is more common than the average person realizes. The Website, Retraction Watch, does a thriving business; indeed,  it seems they can hardly keep up.

And that is just what is blatant enough to be caught. The (sincere but still destructive) misuse of statistics is a huge problem, especially issues of effect size, such as those relating to the difference between statistical significance and practical significance. One scientist quoted widely by pro-Damore journalists, Deborah Soh, wrote:

This includes a study that analyzed the exact same brain data from the original study and found that the sex of a given brain could be correctly identified with 69-per-cent to 77-per-cent accuracy.

Use your common sense, folks! I don’t want to go off on a tangent here on the details of the studies and counter-studies, not to mention how much of that gender difference itself is due to Nurture (thinking produces furrows in the brain etc.), but what in the world does this “brain gender identifiability” have to do with software development? Are the brain measurement differences seen by the researchers large enough to make any kind of practical difference in programming talent? Soh is making a huge leap here. Similar statements can be made for the various distributional differences cited by Damore in his essay.

A related matter is the common study flaw I mentioned in my blog post, in which rats with virtually no variation in Nurture are tested and found — surprise, surprise — to indicate that Nature is much more important than Nurture.

It is thus no wonder that research in the Nature vs. Nurture issue has been riddled with the academic equivalent of “he said, she said” disputes, with many studies supporting each side of the issue. Obviously Damore is wrong to simply choose studies supporting his side, and even more wrong to make statements claiming that engineers such as he rely only on facts, and worse still for couching his opinions in ideological terms. I dare say that Damore and his supporters live in their own Echo Chamber, just like the one he rightly pointed out rules at Google.

I stated my belief that in intellectual activity such as software development, Nurture has far more effect than Nature. (Nurture, of course, is meant as a metaphor for all inputs a person receives, both while growing up and later on.) I wrote that this belief has been formed by long years of observation of students, children and so on. Note, though, that I simply stated it as a belief; I did not offer proof, and am in fact skeptical that it is possible to firmly prove or disprove such a thing.

For some readers, the assumption of a major Nature effect was so ingrained that they took statements I made in my post citing certain differences between the genders as a contradiction to my dismissal of Nature. They didn’t notice that I was attributing the differences I cited to Nurture.

Others referred me to such-and-such an author or so-and-so a YouTube video, saying that I then would see the light. Sorry, no light emerged for me. Could be my own failure, but my view, again, is that their views are so ingrained that they saw certain things as “obvious” which might be seen by some others as vague, inconclusive or plain old wrong.

One reader whom I particularly respect brought up the subject of competitive chess, in which men dominate, and which he seemed to show that Nature plays a major role in chess-playing ability. What he didn’t realize that I was planning to cite chess as yet another reason for my belief that Nurture is the key factor, not Nature.

I have been fascinated, indeed inspired, by the story of “The Queen of Katwe,” Phiona Mutesi, an African girl who grew up in such poverty that the word abject would be a gross understatement, but who became an amazing chess prodigy. Nurture here was not having a grandmaster coach and so on, just an extremely intense level of single-minded determinedness to beat everyone in sight. She dreams chess openings and end games like the top programmers dream about compilers and OOP languages. I submit that these traits of hers came from Nurture, especially the excruciating poverty and injustice in which she grew up, and her mother’s heroic efforts to keep the family together after the passing of her husband.

Again, Ms. Mutesi’s story is just one of many, many incidents that have shaped my views, hardly proof. On the contrary, some would insist that she has a Chess Gene. But really, does that sound plausible to you? Well, not to me.

Just how far does the Damore crowd want to take this gene stuff, anyway? Why stop at saying that Linus Torvalds (inventor of Linux) has a Programming Gene? Maybe what he really has is an Open Source Operating System Kernel Gene? And that fellow Einstein must have really won the genetics sweepstakes, having been blessed not with a Physics Gene but with both a Relativity Gene and Photoelectric Effect Gene, eh? Economics? Well, obviously John Keynes was born with a Deficit Spending Gene while Arthur Laffer possessed a Trickle Down Gene.

Pardon my sarcasm (second post in a row now). But I hope I might have planted here at least a seed of doubt in the Damore crowd as to the validity of his arguments.

Advertisement

38 thoughts on “Postscript on Google and James Damore

  1. >had been forced to fire him for legal reasons

    balderdash

    >I hope I might have planted here at least
    >a seed of doubt in the Damore crowd as to the validity of his arguments

    I think you’re still not a biologist.

    Like

    • As I understand it, the legal situation is as follows: Google uses peer review for determining raises and promotions. A female teammate of Damore’s who got a bad review could sue the company for allowing a “biased” person like him to participate, not only with his own review of her but also for poisoning the environment, leading to bad reviews from the other workers.

      Like

      • That’s good to know. I think it’s much more devious to have his colleagues talk about how women are more prone to anxiety, less passionate about coding, etc. in the name of proven “scientific facts” or “an open discussion”, insidiously planting the idea that women are liability. Not to mention his observations, even if they’re true, don’t apply to exceptional people who work at google and do not represent the population traits.

        Like

  2. Norm, I am afraid you are mistaken about Phiona Mutesi, the subject of “The Queen of Katwe.” Rather than being an “amazing chess prodigy,” she is in fact an exceedingly weak player. Her current FIDE rating is a mere 1628, which is typical of a mediocre club player. My freshman roommate in college, with a rating over 2000, was much stronger, and true prodigies are vastly stronger still. I suppose it is kind of impressive that she managed to reach even that level with her background, but it’s really no kind of evidence for anything.

    Like

      • Well the fact that you used the phrase “amazing chess prodigy” led me to believe you considered Mutesi to be a strong player, which she most definitely is not. The fact that they made a movie about her probably gave a lot of people that impression. The movie seemed to disappear without much trace though. Did you see it? I’d be curious to know whether the movie represented her playing ability honestly.

        However if you are interested in arguments for Nurture over Nature, you should be looking at the story of the Polgar sisters. I suspect the girls were all highly intelligent to begin with, and that the training would not have worked with most people, but still…

        Like

        • I saw the movie and read the book. Oddly, the movie is less frank about how truly awful a level of poverty Mutesi grew up in. By the way, I checked after reading your post, and the Web says that a one point Mutesi’s rating was 2000. Not the next Kasparov, as I said, but better than the 1600 you cited.

          Like

          • Could you give me the link to the 2000 rating? Technically a Woman Candidate Master is supposed to have achieved that rating at some point, but it appears that Mutesi was given the title (which I hear is a bit of a junk title anyway) without ever coming close to that level. However if I’m wrong I’d be interested in knowing. It would be quite surprising if someone that young dropped nearly 400 points in such a short time!

            Like

          • I really don’t know any more than what we’ve discussed. It could well be that her situation has been overpromoted. But I find it extremely inspiring anyway.

            Like

  3. It’s interesting to note that Einstein’s brain had some abnormalities — size and shape. We don’t have enough “census” type information to make broad, sweeping conclusions, though, about what is truly “normal” and how much variation exists across race, gender, class, etc.

    What we do know is that it’s much easier for hiring managers, HR types, pundits, and know-it-alls in science, media, and politics to pick a model (Nature or Nurture) and run with it, than for them to admit they really don’t know. Imagine an HR pro admitting publicly, “We just don’t know what makes a great mind, or a great employee, until after we have hired them and given them full run of the place.”

    The best answer is always to treat people as individuals, not members of a particular sub-group, gender, class, whatever. That means having to spend the time and the resources to actually understand a person before hiring them. OUCH! Not what HR has in mind; it’s much more convenient to choose a handful of schools, a couple of ethnicities, a few majors, and some PC mumbo-jumbo, and build a model and just turn the crank. “Plug and Chug” HR.

    As you noted, Norm, personal desire for excellence trumps everything else, anyway. You can’t really get that from a brain scan or a set of C.V. “bullet points”.

    Like

  4. Norm, any discussion of biology, etc. deflects from the most important feature of this event: Google’s violations of EEO law.

    1. EEO (Title VII of the 1964 Civil Rights Act) is a law which covers corporations. In other words, only corporations can “EEO”. Title VII is part of federal employment law.
    2. All US workers can haul any corp into court under this law when the EEO office makes sure the worker’s charges will NOT be a nuisance lawsuit and so issues a “right to sue” letter. When the company is large, like Google, most employment lawyers will represent the worker on contingency.

    3. Google is well-known for age discrimination.
    4. Google is in the midst of a court battle for gender discrimination in wages.
    5. Google offers jobs to foreign nationals 6 months before they need them in the job. This is thru the H-1b program. Until US workers are routinely offered jobs 6 months ahead of time, Google is committing national origin discrimination.
    6. Google is likely committing race discrimination in hiring, etc.
    7. Google’s memo policies, etc, allowed corporate memo’s to be penned advocating that Google continue to violate EEO laws.

    For those who don’t understand the impact of #7, imagine memo’s being penned advocating violating EPA laws, finance laws, etc. Google’s climate, allowing memos that support Google’s EEO violations is a smoking gun for EEO lawsuits!

    Google needs to clean house. They need to train their workers on how to follow federal employment laws instead of how to discuss them. I agree with you, Damore should have been trained the hell out of, along with all other employees.

    But then, doing so would mean that Google has shifted to following EEO law.

    However, given the brewing gender discrimination suit that hasn’t even been filed (60 workers), coupled with the age class action (300 workers), it’s very, very likely that Google’s discriminatory days are numbered.

    Like

    • Are you sure that most employment lawyers will take cases on a contingency fee basis? My impression is the opposite. Employment discrimination cases are quite difficult to win, so the lawyers are reluctant to go contingency.

      What you ae talking about is discrimination on the basis of immigration status, not national origin. Actually, foreign-born engineers and programmers face the same problems as U.S. natives once they get green cards.

      Like

      • Norm, yes lawyers will when the company is large the harm to US workers is broad and the damages deep. In other words, lawyers must weigh the risk of working without pay for many years. They can only justify this when the settlement or lawsuit is in the millions of dollars. Since tech workers have high wages, the damages are also high. However, some will take on EEO suits on contingency for non-financial reasons too.

        Before the mid-80s discrimination cases weren’t THAT difficult to win because companies continued to discriminate despite the passage of these employment laws. Since the tech industry operates like they are living in the 1950s (when it was legal for private companies to block qualified candidates based on race, gender, etc) these lawsuits are easier to win than, say, filing against universities in the early 2000s.

        Norm, immigration status/citizenship was not a protected class until the passage of the 1986 Immigration Act. These types of discrimination fall under this law.–not the Civil Rights Act. It is also quite weak. For instance, the company must show intent; such as “H-1b Only” want ads or an email/memo stating “we won’t hire any citizens or green card holders for this position.” Even when all workers are visa holders and they never interview other statuses, until the company shows intent the workers won’t win under the 1986 Immigration Act.

        The ’64 Civil Rights act (Title VII) does have national origin as a protected class. It’s being used in the Infosys & Tata EEO class lawsuits. This law covers the impact of hiring & employment practices–with or without intent. For example, when a company restricts the applicant pool for interviews by only selecting South Asians, they will be fined under this law. So a company exec can say, “we didn’t mean to discriminate against others” but it does NOT matter what they “meant”. Only outcomes matter.

        EEO looks at outcomes–not inputs like the ’86 Immigration Act. In other words, under EEO a court can look at the totality of employment actions to discern a “pattern or practice” of discrimination.

        One point not lost on the lawyers is that once an EEO suit is filed, the company must essentially prove a negative–that they don’t discriminate and ARE following all EEO policies. They must hand over all docs, stats, emails that the plaintiffs (their enemy) ask for. As such, it’s a bit like handing your enemy ammunition to use against you. This is why Google is refusing to comply with the DOL’s request for evidence.

        Like

        • Donna very glad to see you in this forum, I still follow your website.

          I’m not sure why anyone is still debating Diversity at Google, other than pictures on the webpage (I find the term “Greyglers” very offensive by the way), there’s very little evidence suggesting Google is interested in Diversity.  There is however, strong evidence suggesting Google is very interested in replacing American citizens of all classes; minorities are most vulnerable, with cheap indentured labor in the name of Diversity.  Google demographics don’t support the assertion of a diverse environment (in any form).  Google already has a disproportionately high number of Asian population Google 35% vs 4.8% per 2010 US census (7 TIMES HIGHER THAN EXPECTED), but Google execs want more “Alphabet executive chairman Eric Schmidt made it clear. ‘The stupidest policy in the entire American political system was the limit on H-1B visas.’”  A Google executive advocating for more H1B?  This even while American minorities are already shockingly under-represented at Google (Intel, Facebook…), blacks Google 2% vs US 12% (6 TIMES LESS than expected) and Hispanic Google 4% vs US 16% (4 TIMES LESS than expected), even the white population is under represented to a degree US population 72% vs Google 56%, because Google executives overlook American citizens in favor of cheap indentured labor, mainly from India.   Consider that 70% of the H1B visa go to South Asians, remove the expected ~5% and that suggests that 21% of the population at Google could be South Asian (23 TIMES HIGHER than expected population for Indian Americans in 2010).  Add to this the dirty little secret that Google, Intel, Facebook and others don’t count contract employees from Infosys, HCL or Tata in their demographics and you will find the number of south Asians is much higher. So instead of investing in our under-represented minorities they import more guest workers creating a less diverse environment.

          Like

          • I believe that Google sincerely wants diversity. But I equally believe that (a) some of the top people secretly agree with Damore, (b) they sincerely (but wrongly) think the “diverse” pool just isn’t out there to hire from, and (c) they suffer from crippling unconscious biases.

            I’ve mentioned before the case of a young black man from an absolutely top U.S. university whom Google hired as a summer intern. Nice, except they shunted him into a nontechnical job.

            Like

  5. I’m ambivalent. I look at individuals as…. individuals. Nature plays done offset in larger populations, but to what degree is to be battled out among those who study biology and its many offshoots.

    I’m missing where Damore took a negative path in mentioning alternatives in promoting Comp Sci to females. The memo was internal, the intent appears to be a wake up call to Google after suffering through a highly politicized “re-education” session.

    For those who actually read the entire document, I’ve seen the negative reactions run from “He’s a White Male supremacist genetic nutjob!” to a plain old political backlash of “how dare he challenge Chairman Mao’s re-education camp sessions!”

    I’ve visited with the Googlies three times, and the passive aggressive undercurrent was palpable. The culture gave me the creeps, so I broke off all contact with them.

    Regardless of which side you take, my advice to all who may consider working for Google is to turn to the NSA instead. They have more mentally healthy employees, and less propensity to spy on Americans.

    Like

  6. For a thought experiment – how about taking the example of musical talent? Do you agree that there is some significant degree of heritability in that? Most experts in the field agree that there is (although that is not “proof”). Does not a Mozart have to be “born”? Yes, he had a musical tutor father, but did his “nuture” prior to age six exclusively create the musical genius? And yes, there is significant evidence (including separated twin studies) that intelligence generally (not isolated down to “mathematics” or “programming” talent) has a significant heritable component. So why should there not be some slight measurable differences between different genetic groups in aggregate (allowing that does not predict for the individual)?

    Like

    • All I can do is beg you to read my posts carefully. 😦

      I never said that no trait is genetic, and even for programming all I really said was that Nurture has far more impact than Nature.

      I do believe that playing music likely has a significant Nature component. But maybe not for composing music.

      As to the identical twin studies, I’ve read many of them, and indeed had them in mind when I made my negative remarks about scientists.

      As to your comment on “slight measurable differences,” again I implore you to read my post.

      Like

  7. So your argument basically boils down to “Nurture has far more impact than Nature” – with emphasis on the very difficult to quantify (and hence to prove/disprove) matter of degree? Then that is not saying very much at all.
    I don’t think Mr. Damore was saying nurture had no impact.

    Like

    • The boild-down version:

      1. In something like programming, Nurture has far more impact than Nature.

      2. Even if the “significant differences” found by scientists are genuine, one has NO idea how much impact they have on programming etc.

      3. Damore was highly selective in his choice of studies, and also suffers from confirmation bias.

      Like

  8. And your comment about twin studies relating to your “negative remarks about scientists” – I gather relates to their abuse of statistical methods?
    I think perhaps you are conflating the identical versus fraternal twin studies as opposed to separated [identical] twin studies. The former does rely on statistical tests to tease out the additive contribution of genetics over environment, which is often a relatively small effect. However the separated twin studies only have to show that the twins are more correlated (usually much more so – beyond statistical question) than any two random individuals. [The separation studies do leave open the question of “in utero” environmental effect.]

    Like

    • Yes, there is the question of the prenatal care the twins (fraternal or identical) have in common, as well as the fact that twins raised apart tend to be raised in the same socioeconomic class. But there is much more than that, though subtle. One point I haven’t made so far, due to lack of time/space, is looks, i.e. whether someone is considered good-looking or not. Supposedly teachers and even parents favor better-looking kids, resulting in their getting more or better Nurture and also resulting in higher levels of confidence and so on.

      Life is complex. 🙂

      Like

      • Valid points about “looks” and “class” – but separated twin studies are so interesting because they often wind up so similar in many respects despite very different upbringings.

        Like

          • Yes, I heard the points about looks and class before, but the separated twin effects are so very strong that it is quite hard to attribute them to that. In the case of looks it is impossible of course, to exclude the effect [and the in utero effect as well]. In the case of class it should be possible, but unfortunately separated twin numbers are so small that controlling for the effect is not readily possible.

            Like

  9. The fundamental error in google boy’s theory is that any female CS has to pass Calculus I, II, III and another 3 or 4 upper level math courses to graduate. I don’t know of any American educated man or woman CS that has/can BS their way through that, although I do question the credentials of some foreign educated workers. If google boy is claiming that even after going through a difficult curriculum women are still deficient. Then, you have to conclude he is making up reasons to discriminate.

    As a female CS, I have worked in places where women are respected and places where they are not. In every case, the pattern of behavior was set at the highest levels of the organization.

    If google boy wants to follow in the footsteps of Charles Murray, that’s fine, but he should do it on his own time.

    Like

    • Funny that he didn’t even majored in CS so he probably didn’t have to pass Calculus 1, 2, 3. He even admitted he’s not passionate about coding. Academically and motivationally he’s under-qualified to judge his female colleagues.

      Like

  10. One of Damore’s assumptions was that the employer’s culture was ideal or maximally efficient, and therefore it was legitimate to expect applicants to match that culture.

    The best managers I had were the ones who could made the best use of the talents in the work group.

    I worked in manufacturing under the “quality culture,” where I saw the power of adjusting the system to match the people. That is, the *system* was the object of constant improvement, not the people (so much). Of course different workers have different temperaments, interests, and motivations. A good manager can create a work environment where you optimize the organization around the people and products.

    Fundamentally, that is the power of diversity. If you make best use of different viewpoints, you get better outcomes. Women, just like any group, have valuable life experience to bring to problem-solving.

    Like

  11. Business Insider used to publish articles on the interview process at google. They thought questions like how many golf balls fit into a school bus where cool. I would have preferred to discuss how to make search more efficient and other ways to monetize it. But what do I know, I’m just a woman.

    Like

    • Google did an introspective study on their hiring practices, and as a result decided not to ask “golf ball” questions. But from what I’ve heard, they still do. I actually think such questions can be useful, but I think the “puzzle” problems are not.

      Like

  12. I risk sounding bonkers here, this is just my G.K. I think Science is too young to have this debate about intellect being nature or nurture. Right now, we don’t have integration of physics, biology & philosophy to properly answer many fundamental questions about reality. Logic is incomplete, as per Godel’s/Turing’s theorems. So, there is something more fundamental/abstract than logic (say X) from whose ability mathematics & arts both arises (same way) & allows us to dwell in meaning. Unless we understand the biological basis for X, we can’t answer, in scientific terms, why some people excel in specific fields (like Einstein or Mozart). If women can excel (dominate) in authoring substantive novels, journalism, library science etc., then, why not software development or philosophy? We just don’t know the Science yet & we’d otherwise be throwing around meaningless correlations.

    You may be tempted to ask, if we should’ve quotas for poor candidates & quotas for women then why not treat high skilled immigration the same way? My answer to this question is quite simple if you think that immigration today is solely driven by need for a better life, then levels of immigration (women/men, STEM/non-STEM) are dependent on the state of the ‘American dream’. The issue of equality & quotas for women is independent of the state of the ‘American dream’.

    As far as James Damore is concerned, isn’t it too coincidental that his comments came just weeks after John McEnroe’s comment about Serena Williams & biology of women? Google has had a great impact on societies everywhere, but, I can’t help but remember one article that my female colleague was reading when I was a recent hire: A 25 billion dollar eigenvector.

    https://www.rose-hulman.edu/~bryan/googleFinalVersionFixed.pdf.

    Since then, I never had an impression that it requires real genius to work at Google. For Google, it wasn’t a question of nature or nurture, but rather of being in the right place at the right time. I don’t understand all the media attention.

    Like

  13. I’m missing something. You (as a hiring manager) play the cards you’re dealt. What does it matter if the cards you were dealt (a pool of candidates, male and female, strong and weak) has its subset of strong-enough candidates not matching the general population by reason of nature or by reason of nurture? “Nature” is immutable, but “Nurture” from conception until the present, is also already written.

    Like

    • I agree. But remember, my other major point was whether the hiring managers are accurately assessing what is already written, which I suspect they are not.

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s