Author: Megan Molteni and Adam Rogers
The Actual Science of James Damore’s Google Memo
In early August, a Google engineer named James Damore posted a document titled “Google’s Ideological Echo Chamber” to an internal online discussion group. His memo was a calm attempt to point out all the ways Google has gone wrong in making gender representation among its employees a corporate priority. And then, on August 5, the memo jumped the fence. Nobody else was calm about it.
It wasn’t a screed or a rant, but, judging by his document, Damore clearly feels that some basic truths are getting ignored—silenced, even—by Google’s bosses. So in response, the engineer adopted a methodology at the core of Google’s culture: He went to look at the data. “Google’s Ideological Echo Chamber” wants to be a discussion of ideas about diversity through solid, ineluctable science.
The core arguments run to this tune: Men and women have psychological differences that are a result of their underlying biology. Those differences make them differently suited to and interested in the work that is core to Google. Yet Google as a company is trying to create a technical, engineering, and leadership workforce with greater numbers of women than these differences can sustain, and it’s hurting the company.
Damore further says that anyone who tries to talk about that paradox gets silenced—which runs counter to Google’s stated goal of valuing and being friendly to difference. And, maybe helping make his point a little, last Monday Google fired him. Damore is now on a media tour, saying he was fired illegally for speaking truth to power. Hashtag Fired4Truth!
The problem is, the science in Damore’s memo is still very much in play, and his analysis of its implications is at best politically naive and at worst dangerous. The memo is a species of discourse peculiar to politically polarized times: cherry-picking scientific evidence to support a preexisting point of view. It’s an exercise not in rational argument but in rhetorical point scoring. And a careful walk through the science proves it.
The Incoherency Problem
Psychology as a field has been trying to figure out the differences between men and women, if any, for more than a century—paging Dr. Freud, as the saying goes. The results of these efforts are ambiguous. And psychologists are still working on it.
The science of difference is a mushball, and trying to understand differences among populations only makes it messier. Every cognitive or personality trait will have a wide distribution among a given population—sex, ethnicity, nationality, age, whatever—and those distributions may only vary slightly. Which means huge chunks of the population may overlap. For any given trait, men may be more different from each other than from women, let’s say.
That said, Damore’s assertion that men and women think different is actually pretty uncontroversial, and he cites a paper to back it up, from a team lead by David Schmitt, a psychologist at Bradley University in Illinois and director of the International Sexuality Description Project. The 2008 article, “Why Can’t a Man Be More Like a Woman? Sex Difference in Big Five Personality Traits Across 55 Cultures,” does indeed seem to show that women rate higher than men in neuroticism, extraversion, agreeableness, and conscientiousness.
As always, the issue is the extent of the difference (and what causes it—more on that in a bit). Also, as Damore himself notes: Google hires individuals, not populations.
Damore argues that greater extraversion and agreeableness, on the whole, would make it harder for women to negotiate and stake out leadership positions in an organization, and that higher neuroticism would naturally lead to fewer women in high-stress jobs. The first-order criticism here is easy: Damore oversells the difference cited in the paper. As Schmitt tells WIRED via email, “These sex differences in neuroticism are not very large, with biological sex perhaps accounting for only 10 percent of the variance.” The other 90 percent, in other words, are the result of individual variation, environment, and upbringing.
A larger problem, though, is measuring the differences in the first place. Personality traits are nebulous, qualitative things, and psychologists still have a lot of different—often conflicting or contradictory—ways to measure them. In fact, the social sciences are rife with these kinds of disagreements, what sociologist Duncan Watts has called an “incoherency problem.” Very smart people studying the same things collect related, overlapping data and then say that data proves wildly different hypotheses, or fits into divergent theoretical frameworks. The incoherency problem makes it hard to know what social science is valid in a given situation.
The impulse to apply those theories to explain human behavior is as strong as it is misguided. Women as a group score higher on neuroticism in Schmitt’s meta-analysis, sure, but he doesn’t buy that you can predict the population-level effects of that difference. “It is unclear to me that this sex difference would play a role in success within the Google workplace (in particular, not being able to handle stresses of leadership in the workplace. That’s a huge stretch to me),” writes Schmitt. So, yes, that’s the researcher Damore cites disagreeing with Damore.
Damore does this over and over again, holding up social science that tries to quantify human variation to support his view of the world. In general, he notes, women prefer to work with people and men prefer to work with things—the implication being that Google is a more thing-oriented workplace, so it just makes sense that fewer women would want to work there. Again, the central assertion here is fairly uncontroversial. “On average—and I emphasize that, on average—men are more interested in thing-oriented occupations and fields, and that difference is actually quite large,” says Richard Lippa, a psychologist at Cal State Fullerton and another of the researchers who Damore cites.
But trying to use that data to explain gender disparities in the workplace is irrelevant at best. “I would assume that women in technical positions at Google are more thing-oriented than the average woman,” Lippa says. “But then an interesting question is, are they more thing-oriented than the average male Google employee? I don’t know the answer to that.”
Semantics aren’t helping here. Is coding a thing- or people-oriented job? What about when you do it in a corporation with 72,000 people? When you’re managing a team of engineers? When you’re trying to marshal support for your proposed expenditure of person-hours versus someone else’s? Which is more thing-oriented, deep neural networks or database optimization?
And maybe the most important question: How useful are psychological studies of the general population when you’re talking about Googlers?
Nature vs. Nurture
Damore essentially forecloses the possibility of changing sex roles and representation at Google—or anywhere, really—by asserting that not only are the differences between men and women significant but that they are at least in part intrinsic. Damore doesn’t assert that biology is the only factor in play, and no scientist does either. But how important biology is to psychology is—again—in heavy dispute.
Here’s Damore’s take: “On average, men and women biologically differ in many ways.”
Nothing to argue about here. If men and women didn’t differ biologically, it would make sexual reproduction very difficult indeed. Also, men and women differ in height (on average), bone mass (on average), and fat, muscle, and body hair distribution (on average). No one thinks those differences are socially constructed.
Damore, though, is saying that differences in cognitive or personality traits—if they exist at all—have both social and biological origins. And those biological origins, he says, are exactly what scientists would predict from an evolutionary perspective.
Evolutionary psychology and its forebear, sociobiology, are themselves problematic fields. Two decades ago evo-psych was all the rage. It’s essential argument: Males and females across species have faced different kinds of pressures on their ability to successfully reproduce—the mechanism, simplistically, through which evolution operates. Those pressures lead to different mating strategies for males and females, which in turn show up as biological and psychological differences—distinctions present in men and women today.
The problem with that set of logical inferences is that it provides a convenient excuse to paint a veneer of shaky science onto “me Tarzan, you Jane” stereotypes. It’s the scientific equivalent of a lazy stand-up comedian joking about how all men dance like this—the idea that nature hardwires our differences. In fact, evolutionary biologists today race to point out that the nature-versus-nurture dichotomy is outdated. No serious scientist finds it to be a credible model.
In 2005, Lawrence Summers, then president of Harvard, suggested publicly that women might not have as much “innate ability” as men to succeed in academic disciplines that require advanced mathematical abilities. In response, psychologists got together to assess more than 100 years of work and present a consensus statement about whether Summers was right. They concluded that a wide range of sociocultural forces contribute to sex difference in STEM achievement and ability, including family, neighborhood, school influences, training experiences, cultural practices, and, yes, some biological factors.
When it comes to brain biology in particular, the authors wrote that “experience alters brain structures and functioning, so causal statements about brain differences and success in math and science are circular.” Most researchers today point to data that shows cognitive traits differ slightly on average between the sexes, but they change throughout an individual’s lifetime, influenced by a mix of genetic, epigenetic, and environmental (including social) factors.
From birth, boys and girls receive different, gender-specific treatment, which can enhance or inhibit any innate differences. That certainly has an effect on the findings of psychology. The gap between girls and boys who say they want to go into the sciences is much more informed by stereotypes—on a survey of half a million people, 70 percent associated math with males—and cultural norms than by intrinsic ability. “From infancy, boys get footballs and girls get dolls, so is it that surprising? We’ve been socializing them. It doesn’t mean there’s anything innate,” says Janet Hyde, director of the Center for Research on Gender and Women at the University of Wisconsin.
All these things change as culture changes. In 1990, Hyde published a meta-analysis on sex differences in mathematical performance among high school students and found significant deficits in girls’ abilities. When she did the same analysis in 2008, the difference had disappeared. In the 1980s, “girls in high school didn’t take as many years of math as boys did,” Hyde says. “Today that gap in course taking has closed. Girls take as many classes as boys do, and they’re scoring as well. What we once thought was a serious difference has disappeared.”
And just as culture moves on, so too does biology. “The brain can change a lot in a matter of weeks,” says Diane Halpern, an author on that post-Summers study and of one of the central textbooks on cognitive sex differences. “That’s why we send children to school. There are areas where, on average, women excel and, on average, men excel, but everyone gets better with education. But it means we cannot know the influence of environmental versus biological variables, even at very young ages.”
In other words, the science on math and science abilities says differences between sexes depend much more on external factors than sex in and of itself. And those external factors and their results can change over time.
This is critical, because most of Damore’s memo seems to be talking about preferences, which is to say, rather than innate skill he means what women would rather be doing versus what men would rather be doing. In fact, one recurring finding in sex difference research is that in cultures seen as more egalitarian, differences in preferences between men and women become more pronounced. With more opportunity, says one hypothesis, men and women are more likely to follow their respective blisses.
So when Damore does juke from preferences to abilities, it looks a little sneaky. Here’s what he writes: “I’m simply stating that the distribution of preferences and abilities of men and women may differ in part due to biological causes and that these differences may explain why we don’t have equal representation of women in tech and leadership,” he writes. Making the leap from personality differences to achievement differences would require citing at least some of the well-studied body of work we’ve mentioned here, which Damore ignored.
With the next pivot, the memo gets more pernicious. Damore switches—again, subtly—from effects to causes. His interpretation of the science around preference and ability is arguable; on causation, though, he’s even rockier. According to Damore (and a lot of research), the biological factor that connects sex to cognitive abilities and personality traits is prenatal exposure to testosterone.
Of all the high-stakes claims in sex-difference research, none is more important or more popular than the idea that hormones in the womb help give people stereotypically masculine or feminine interests. While they’re developing, males get a bigger dose of testosterone. “Among social psychologists there’s a consensus that prenatal testosterone does affect a lot of personality traits, in particular one’s interest in people versus things,” Damore said in an interview last week with Bloomberg’s Emily Chang. He also said it to pro-Trump YouTuber Stefan Molyneux, adding that hormonal exposure “explains a lot of differences in career choice.”
Damore is probably wrong about this too. The most consistent findings linking prenatal testosterone to sex-linked behaviors come from about a dozen studies examining toy preferences among girls with a condition known as congenital adrenal hyperplasia, which causes the overproduction of sex hormones, including testosterone. CAH-affected girls tend to be less interested in dolls (substituting for people) and more interested in toys like trucks (things).
But children with CAH have other variables. They’re often born with ambiguous genitalia and other grave medical conditions, and therefore have unusual rearing experiences. To get around this socialization issue, researchers from Emory University gave toys to young rhesus monkeys. When they saw that females preferred plush dolls and males preferred trucks, they concluded that these tendencies must be hard-wired into each sex.
Squint hard at this result, because it presumes that juvenile rhesus monkeys see stuffed animals as monkeylike but “wheeled toys” as thinglike. But why would a monkey see a plush turtle as akin to self? And how would it know what a truck was or was not? Also: The male monkeys played with trucks. The females chose between the two about equally. The logic here walks a twisted path across the floor of the uncanny valley.
Still, most hormone researchers agree that these differences are real. But that they’re directly linked to prenatal testosterone? Not so much. And to differences in career choice? “There’s 100 percent no consensus on that,” says Justin Carré, a psychologist at Nipissing University in Ontario. “The human literature on early androgen exposure is really very messy.”
Appeal to Science
Damore needs scientific consensus to make his case—not just because of confirmation bias but because the memo goes on to argue that the left is just as guilty as the right when it comes to science denialism. He equates conservative tendencies to reject climate change and evolution (theories with an overwhelming scientific consensus behind them) with liberal refusals to accept differences in personality traits between the sexes and—in a quiet racist dog whistle—IQ, where the evidence is far, far weaker.
Climbing to an even higher altitude, though, we might ask another question about Damore’s appeal to science: So what? Which is to say, what are we to do with not just the conclusions of the memo but also its implications? Damore is hardly the first person to use science to justify social norms or political preferences. Science has, too often in human history, been a tool for literal dehumanization as a rationale for oppression. It happened to people of African descent in America; to the poor of the Victorian era; to women in the years leading up to suffrage; and to Jews, people of nonbinary gender, Roma, people with disabilities, and so on in Nazi Germany. Historians try to wall off those ideas now—eugenics, phrenology, social Darwinism—but each, in its day, was just science.
With hindsight you can see that those pursuits weren’t science, and you can aim those 20/20 lenses at Damore too. What he’s advocating is scientism—using undercooked research as coverage for answering oppression with a shrug.
In that context, social science’s incoherency problem becomes disastrous. Throw the most red-state conservative physicist you can find into a room with a pinko-commie physicist and then toss in the latest data from the Large Hadron Collider. Mostly, the physicists will agree on which subatomic particles they can or can’t find. But even if you buy the research on psychological sex differences, the work in their biological or evolutionary basis is far from finished—leaving people free to cherry-pick results ready to mix into a manifesto. Just add outrage.
Science must inform policy—social, corporate, whatever. The more solid the science, the more it can inform. (Why, hello, climate change data—you are terrifyingly real.) But when it comes to sex differences, Google—or any organization, really—will understandably want to create an environment where people feel secure, safe, and empowered to do their best work. It’s good ethics and good business. That’s what Damore seems to see as an overly politically correct culture that stifles dissent.
Damore’s dissent, stripped of its shaky scientism, isn’t a serious conversation about human difference. It’s an attempt to make permanent a power dynamic that shouldn’t exist in the first place. If Google was, for Damore, an echo chamber, that’s because his was the only voice he was really willing to hear.