A Measured, but Failed, Response to Damore

The opposing replies to James Damore’s memo have by and large been pretty terrible. They strawman, they make sweeping and inaccrate claims about the state of the science, they show little engagement with the content of the memo, and they’re negative and emotional to boot. Computer Science professor Cynthia Lee wants to do better. She wants to be one of the ones who argues against Damore without going going over the top or playing dirty – so she posted a response on Vox. Does she succeed?

Not really. Here are the problems I see with her effort.

At the outset, it must be conceded that, despite what some of the commentary has implied, the manifesto is not an unhinged rant.

I’m happy to see this here, but I think it’s inadequate. Because you see, it isn’t just that Damore’s memo was unfairly characterized as "an unhinged rant," it’s that most of the responses to him you read online actually are "unhined rants." That needs a mention here. She does acknowledge it – but only to euphemize it away:

Many defenders seem genuinely baffled that a document that works so hard to appear dispassionate and reasonable could provoke such an emotional response.

The response is not merely "emotional," it’s childish. It’s a raft of foot-stomping temper tantrums involving a lot of outright misrepresentations of Damore’s memo. "Emotional" is too broad a word for the fairly narrow range of responses we’ve gotten, and so this reads like a deliberate dodge.

Much of the science it cites, too, has at least some grounding in peer-reviewed research, even if the author’s conclusions are not justified by the findings, failing to adequately account for sociological and other factors.

This is simply not true. Now, let’s be careful here: science is a process of findings and counterfindings – reports and corrections. So, it’s often possible to cherry-pick studies that support your political bias, and it’s important for people reading scientific literature to guard against confirmation bias in themselves when evaluating studies. Damore may well be guilty of this. But it is not true that Damore’s conclusions are not justified by the findings, nor that the literature in quesiton argues for a greater role for socialization than Damore acknowledges. Damore’s point, in any case, was not that biology is primary over socialization – it was that Google was possibly assuming a greater role for "sociological and other factors" than the literature supports. Lee is missng the context: Damore doesn’t have to explain the socialization part of it because Google is already solid there – too solid. He’s not trying to talk Google out of considering sociological factors; he’s trying to talk them into looking at anything else.

"Note that these are just average differences," the manifesto reiterates, soothingly, "and there’s overlap between men and women." Here again, this studious dispassion and showy air of reasonableness create cover for the memo’s defenders. They have been vociferously arguing online that women at Google are not “average” and so they should not be offended by the manifesto’s litany of citations to studies of the “average” woman’s deficiencies.

Calling actual Damore’s actual reaonableness a "showy air of reaonableness" makes for some fun irony: we now know that Lee is not playing fair and can guess that she herself adopts a reasonable tone to appear even-handed while not actually being so. Black pot, meet black kettle! But the real problem here is the lack of counterargument to these "defenders." Leave aside for a moment that Damore’s point was not that women make bad coders or even that Google should hire fewer of them – it was rather that Google was wasting energy on approaches to diversity that are unlikely to work and may undermine the company’s effectiveness. Given research on sex-based personality differences, Google needs to either accept its current level of gender balance or try different methods.

That’s the introduction. Here’s the thesis: Lee is going to tell us why we shouldn’t be surprised there was so much outrage in response.

First up – Fatigue.

It’s important to appreciate the background of endless skepticism that every woman in tech faces, and the resulting exhaustion we feel as the legitimacy of our presence is constantly questioned.

It is important. And that’s why we should be able to discuss freely the idea that affirmative action hires are actively making things worse on this front. If you know that your company is trying desperately to hire people to hit an arbitrary gender balance target, and you suspect that it is doing so at the expense of quality hiring, then you are going to be suspicious of the legitimacy of coworkers who fit the profile of an affirmative action hire. Try a thought experiment. Tech Company A is very worried about its gender balance for legal and public relations reasons. Tech Company B doesn’t give a shit about its gender balance and ruthelessly chases talent. Tech Company A employs more women than Tech Company B, but women work as coders at both companies. At which company do you really think women are more likely to be treated like imposters by their male co-workers? I think the right answer here is "we don’t know; we can spin a rationalist argument either way on that." It could be Company A, because it’s known that Company A makes an effort to hire women and probably hires a lot of false positives. It could be Company B, because male employees at Company B interact with fewer female coders and thus have fewer opportunities to have their stereotypes challenged. Damore wasn’t taking a side on this issue. He was merely saying that answering that question with "Tech Company A" is a thing that should not be censored.

When men in tech listen to the experiences of women in tech, they can come to understand how this manifesto was throwing a match into dry brush in fire season.

Men should listen to the experiences of women in tech. But a lot of these responses aren’t comming from women. More than half of the "unhinged rants" I’ve personally read – yes, annecdote, feel free to correct for selection bias – have been from men. These are not motivated by "fatigue" at having to repeatedly relate their experiences in the office.

Next up – Women’s resistance to the ‘divide and conquer’ strategy

There’s not much to say about this one because it’s just good old-fashioned mendacity. Damore is not pursuing a "divide and conquer" strategy, and he’s not trying to undermine women.

Speaking for myself, it doesn’t matter to me how soothingly a man coos that I’m not like most women, when those coos are accompanied by misogyny against most women. I am a woman. I do not stop being one during the parts of the day when I am practicing my craft.

And here’s where I think we run into some language issues. Damore didn’t write (let alone "coo") that "most women" are unsuited to careers in tech. He was talking about differences in tendencies between group averages. Since a lot of people (including, apparently, Computer Science PhDs) seem to have trouble understanding how overlapping distributions work, perhaps Damore shouldn’t have just left that at the level of an illustration; perhaps he should have spent more time explaining this. When you say that two populations largely overlap, but have different averages, then what you mean is … well, that most people in those populations are actually indistinguishable from each other on the variable measured. So say we’re looking at "neuroticism," one of the personality traits1 Damore suggests account for women’s relative lack of interest in tech fields. With some typical highly-overlapping distributions we’d get two normal distribution curves, one (for women) with a peak (the average) shifted slightly left, and one (for men) with the peak (the average) shifted slightly right. Both curves will have tails at the extreme ends of the spectrum. At the extreme end for "neuroticism" will be … not many people, but more women than men. Meanwhile, the vast majority of people – male and female – will be within the normal range of neuroticism. Meaning: if I pick a point in, say, the first two standard devaitions – where both of these curves (let’s say) overlap, I will be able to infer NOTHING about whether that person is male or female.

Damore understands this, but Lee doesn’t seem to. At least, if Lee understood this, I would have to believe that she could see that what Damore is implying about men and women is that for the clear majority of men and women there will be no discernable sex-linked difference in whether they’re interested in coding. Which leads to the next point:

Third – Damore cites ‘averages’ but Google isn’t ‘average’

Here’s where we know that she’s just completely missed Damore’s argument. Yes, Google isn’t average, and that’s one reason we might expect it to have fewer females than a more normal tech company (if the research Damore cites about tendencies is correct and actually germane). Again, it has to do with how normal distributions work. For the sex-linked cognitive differences Damore cites, there are clear, statitistically significant differences between the sexes, BUT (as Damore himself points out) the differences in average are small and the overlap between the two populations is great. In situations like that, what it means is that the differences get starker the more execptional the population. So, to reiterate, if we’re talking about people within two standard deviations (say) of "normal" for their sex on "neuroticism," then for those people there’s really not much difference between the sexes. Knowing someone is one standard deviation more neurotic than another tells you little to nothing about their sexes. But if someone is high on neuroticism – way out there on the tail – you have a pretty good guess she’s female. So, the fact that Google isn’t average works in exactly the opposite way that Lee wants to say.

Lee gets this to a degree, and she would say that "how average works" is precisely her point:

What do averages have to do with hiring practices at a company that famously hires fewer than one percent of applicants? In the name of the rational empiricism and quantitative rigor that the manifesto holds so dear, shouldn’t we insist that it only cite studies that specifically speak to the tails of the distribution — to the actual pool of women Google draws from?

She’s right about that. Since Google hires from an exceptional pool, the population of women Google is drawing from is unlikely to look like the general population of women across all categories. But it’s also telling that she’s apparently completely uncurious about what that pool actually looks like. Damore’s argument isn’t, after all, about the personality traits of the pool Google has to draw from but rather how the relative size of the female pool to the male one gets to be how it is. She seems to miss that what Damore is doing is asking Google to do precisely the studies she insists he cite in the first place. His whole claim is that they haven’t; that they’ve been running on the unexamined assumption that biology isn’t a factor. Damore’s memo, let’s not forget, doesn’t assert that biological differences explain away the skewed gender ratio at Google, it rather raises the possibility that biology plays a role and calls for active investigation of same. So no, Damore hasn’t missed this point. It IS his point. He’s asking Google to do exactly the studies it’s best-positioned to do and that she claims he’s ignoring. Or rather, not even that much. He’s asking for people to be able to suggest they do these studies without getting fired for the favor.

Fourth – He mentions race in a memo about gender

This is a misinterpretation. He focuses on gender, but he’s talking about Google’s diversity programs in general. So, of course occasional mentions of race belong in here.

Fifth – The author says he’s for diversity but no real-world programs meet his standards

This is both fair and not. On the one hand, it’s obviously unfair in that Damore’s whole point is that Google’s programs are likely misguided and counterproductive. It’s not up to him to support any of them; he’s trying to explain why he’s against them. On the other hand, it’s true that the few suggestions he offers are pretty weak tea.

He does make some recommendations, but they range from impotent (“Make tech and leadership less stressful”) to hopelessly vague (“Allow those exhibiting cooperative behavior to thrive”) to outright hostile (“De-emphasize empathy”).

Yeah, I have to agree with all that. To clarify – I don’t think someone like Damore is a priori responsible for suggesting alternatives. He’s a programmer, and coming up with diversity initiatives has a dedicated department of its own – one that I presume is competent and well-paid. But in so frequently asserting his support for diversity initiatives, he kind of makes himself responsible for suggesting something. The same person who did all that research on gender differences should be able to spare some time for researching alternate approaches to diversity programs. I assume they exist and are not all quota-based. So, for my money, Lee scores a point here.

All told, it’s a pretty weak performance for her, though. She cites irrelevancies ("fatigue"), relies on insinuation, seems to misunderstand the argument to boot. If you throw any of this at her on Twitter she’s happy to use her +1 Shield of Credentialism to ward it off – but letters after your name don’t give you an argument mulligan. I’m glad the tone is improving on this subject. The arguments don’t yet seem to be.

Worst of all, even this more measured essay can’t bring itself to say that whatever Damore’s memo’s flaws, he shouldn’t have been fired over it. Which means that Lee can be as polite as she likes and it doesn’t matter. She’s ultimately still carrying water for deeply intolerant people – people who threaten, bully and silence rather than engaging and discussing. And that makes her one of the bad guys.


  1. It’s important to note here that "neuroticism" is a technical term in the professional personality literature that doesn’t mean exactly what the general public will take it to mean. This is a common problem when shifting between scientific jargon and everyday language.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>