Editor’s note, April 19th, 2019: We’re republishing this story today because of a report at Quartz that — like the infamous memo from former Google employee James Damore — some Microsoft workers are now internally criticizing the company’s diversity policies, suggesting yet again that biological differences may make women less suited for engineering jobs. This story may help explain why that logic is flawed.

In 2017, the idea that biological differences drive social inequality is considered fairly offensive. For the incurious, the taboo around this argument makes it exciting. But unlike people, not all ideas are created equally, and they should not be treated with the same amount of seriousness — especially when those ideas ignore both a broad scientific debate that’s gone on for years and clear evidence that women in tech are excluded more than in other industries.

The idea that women or people of color lack the innate qualities that white men possess to succeed in high-status, elite professions is decades old. And the shape of the argument always looks the same, saying that current social conditions are somehow biologically natural, and that attempts to remedy inequalities are suspect. It is a tired stance in an endless debate, and it says far more about our feelings than it does about science.

This time, those feelings came from a software engineer named James Damore.

After attending a voluntary diversity training session at Google, Damore penned a 10-page memo entitled “Google’s Ideological Echo Chamber.” In it, Damore reaches for science to explain away the gender gap in tech, arguing that female biology is, in part, holding women back. Programs aimed at boosting diversity in tech are “as misguided and biased as mandating increases for women’s representation in the homeless, work-related and violent deaths, prisons, and school dropouts,” he writes.

The manifesto circulated internally for about a month before going viral, drawing condemnation from some quarters, and approval from others. Damore has since been fired and has filed a labor complaint against Google.

In an all-company memo, Google CEO Sundar Pichai said that while “much of what was in that memo is fair to debate,” Damore went too far. “To suggest a group of our colleagues have traits that make them less biologically suited to that work is offensive and not OK,” Pichai wrote.

Damore pushed back on Friday in a Wall Street Journal op-ed, writing that the document was “a reasoned, well-researched, good-faith argument.” He asks: “How did Google, the company that hires the smartest people in the world, become so ideologically driven and intolerant of scientific debate and reasoned argument?”

The answer is that the debate is old and the arguments are well-worn and generally bad. In 2005, Larry Summers, then-president of Harvard University, suggested at an economics conference that women were underrepresented among faculty scientists because of “innate” differences. That year, only four of 32 tenure offers made by Harvard’s Faculty of Arts and Sciences went to women. In 1994, Charles Murray and Richard Herrnstein published The Bell Curve, which focuses on race and IQ, kicking off a controversy that rages to this day. We have been having this argument forever.

In this context, Damore’s memo is particularly sloppy and fundamentally unscientific. The memo doesn’t clearly define what makes for a successful coder. It doesn’t explore what scientists do and don’t know about how biological sex shapes behavior. It doesn’t call on experts to debate. It doesn’t evaluate the strengths and weaknesses of the evidence. And it never once discusses what we know about gender discrimination or its long, sordid history in tech.

The memo’s deficiencies make more sense once you understand that it’s not a real attempt to grapple with facts or reality, but is really just an expression of Damore’s feelings — a reactionary flailing to justify a broken status quo.

What inequality in Silicon Valley looks like

The numbers are stark: there are far more male technical workers in Silicon Valley than there are female technical workers. The National Center for Women and Information Technology places the percentage of women in “professional computing professions” at 26 percent in 2016. In this respect, Google is actually doing much worse than the rest of the industry at large: only 20 percent of its technical employees are women. (Microsoft reports approximately 18 percent, Apple reports 23 percent, and Facebook 18 percent.)

It’s not like gender inequality permeates Silicon Valley all the way through. When counting non-technical workers, companies are much closer to gender parity: Google employs 48 percent women, Microsoft with 39 percent, Apple has 38 percent, and Facebook 60 percent.

But the gender skew tends to get worse in the more powerful levels of the industry. A 2014 report found that only 11 percent of the top 150 Silicon Valley companies have at least one woman in an executive position. In 2017, Crunchbase found that only 17 percent of startups had at least one female founder, and that percentage hasn’t grown since 2012.


Female percentage of select STEM undergraduate degree recipients.
Image: NCWIT, compiled from Department of Education statistics

Meanwhile, only 6 percent of partners in venture capital firms are women, which is down from 11 percent in 1999. This is a downturn with consequences since VC firms with women partners are more likely to invest in companies founded by women.

Women currently represent about 20 percent of bachelor’s degrees in computer science — about the same proportion as women technical workers at Google. That’s all well and good, until you realize that, in 1985, women represented 37 percent of computer science bachelor’s degrees.

Today, women earn more than half of all undergraduate degrees. Within just STEM itself, women earn the majority of bachelor’s degrees in biology, half the bachelor’s degrees in chemistry, and a little under half the bachelor’s degrees in math.

These declines run counter to other trends. Not only are more women going to college and majoring in STEM fields than ever before, but girls are also keeping pace with boys in high school, earning the same number of credits in math and science and even earning slightly higher grades in those subjects. There’s still an achievement gap when it comes to high-stakes tests, like the math portion of the SAT, but even that has shrunk dramatically in the last few decades.

In the 1980s, high-achieving boys on the math portion of the SAT (scoring 700 or above) outnumbered girls 13:1. Now the ratio has diminished to 2.8:1. Something about American math and science education has changed in the last few decades, and the result is more women in STEM than ever before. This is true except when it comes to tech, where female representation is somehow decreasing instead of increasing.

The tech industry is bleeding women workers


Chart showing decline of women in computing positions from 1985 to 2015

Image: NCWIT, compiled from Bureau of Labor statistics

Damore’s memo assumes that the case for diversity is inherently borne of left-wing values, and is therefore somehow suspect. It’s a fallacious assumption that truth lies somewhere in the middle of two poles. But beyond that, it misunderstands the corporate imperative for diversity. It’s a veneer of good public relations, for sure, but it’s also a crass question of money and the American economic advantage in a global economy.

Some projections indicate that the tech industry is about to experience a labor shortage, and it’s not clear where the needed supply of skilled workers is going to come from. Meanwhile, attrition for women is much higher than it is for men in STEM fields — and much more pronouncedly so when it comes to tech. Forty-one percent of women leave tech companies after 10 years of experience, as opposed to 17 percent of men who leave.

“Reducing female attrition by one-quarter would add 220,000 people to the highly qualified [science, engineering, and technology] labor pool,” says a 2008 Harvard Business Review research report.

The technology industry is bleeding experienced workers in a time when labor — particularly experienced labor — is more needed than ever. The primary reason for midcareer attrition is that women feel stalled in their careers: they are promoted too slowly or aren’t able to navigate a transition into management.

Part of the reason is that women in technology aren’t able to tap into the same social connections that men can. Some studies have labeled this “good old boy culture” as shorthand; other studies have simply called it “unfairness.” At its most benign, “good old boy culture” will merely slow a woman’s career trajectory, but at its worst, women experience intimidation, hostility, and harassment.

At some point, the absence of women in technology becomes a self-perpetuating problem. Because women leave, there are few senior women. Without preexisting networks of senior women to mentor junior women, careers stall at a critical midpoint, and the women leave. When looking in from the outside, many young women see a field dominated by men and choose not to enter at all to begin with.

It’s not like good old boy culture doesn’t exist in other fields and industries. Some unfortunate facts about Western history that James Damore neglects to mention (e.g., women not being considered people) have made it so that women have had no choice but to spend the last century or so carving out spaces for themselves in all kinds of male-dominated professions. But the representation of women in tech is dire even compared to science and engineering fields. There is something about tech that’s even worse.

Is the stereotype of computer programming as a masculine activity having a negative effect on women in the industry? Could implicit bias affect women’s performance evaluations and slow their careers over time? Are people like James Damore costing the technology industry $16 billion a year in turnover and attrition? These are possibilities that may be deeply uncomfortable to consider. Inequality is a big systemic problem that, by definition, some people unfairly benefit from. We would all rather be able to attribute our success to our personal characteristics, and wrapping your head around how systems are unintentionally rigged to perpetuate unfairness can be a little taxing.

The numbers point to something cultural about the tech industry that not only keeps women from entering but also drives away the women who manage to join anyway. This is a culture that Kara Swisher describes as “ruled by the equivalent of badly-raised boys who eschewed the kind of discipline and rigor that is the real requirement of success.” That’s the best way to make sense of the numbers, the trends, and the attrition — and it’s the hypothesis that fits with the anecdotes that engineers like Susan Fowler have published about toxic cultures at companies like Uber.

All of this adds up to a perfectly good explanation for the bizarre gender skew in Silicon Valley. It might be a personally discomfiting one to some, but that’s not a good reason to dismiss the long history of women contributing to tech and instead turn to bad science. “It’s almost strange to have to rationally refute it because it is just so wrong,” says tech historian Marie Hicks, a professor at the University of Wisconsin-Madison and author of the book Programmed Inequality: How Britain Discarded Women Technologists and Lost Its Edge in Computing

How did we get here?

Women powered the tech sector during and after World War II, using the codebreaking computer called the Colossus in the UK. Women put together, operated, and troubleshot work on these computers. After the war, the Colossi were destroyed, but women continued to be the main tech workforce. In the US, a team of women programmed the World War II-era ENIAC, or Electronic Numerical Integrator and Computer, although their contributions received little recognition for almost 50 years.

At the time, programming was low paid, not prestigious, and not considered particularly intellectual. But the challenging work didn’t deserve that reputation: Jean Jennings Bartik, one of the ENIAC’s original programmers, said the machine was a “son of a bitch” to program. “It took years and her autobiography coming out for us to know who really did the work because her male managers took credit for it,” Hicks told The Verge in an email.

At first, men didn’t want in on computing. Hicks reports that in 1955, the UK’s Aeronautical Research Department said, “Boys generally prefer laboratory work to computing… this might be due in part to the absence of any recognized career in computing and of any suitable specialist courses.”

Then the power of computing started to become clear — and suddenly this low-paid women’s grunt work started to look appealing to men. The women who had written and tested some of the earliest computer programs were pushed into sales, demonstration, or assistant roles, Hicks writes. Men began taking over the industry — hired into managerial positions because of their people skills. “They’re doing the same work as the men who are replacing them with better salaries and better titles — and the men have less experience,” Hicks says. “It’s crazy-making to say that women aren’t good at technical pursuits, or that there’s any reason other than structural inequality that’s making women stay away from computing work.”

In the 1960s, big employers like the System Development Corporation (SDC) and IBM began to rely heavily on personality profiles and aptitude tests to hire programmers, writes Nathan Ensmenger, a professor at Indiana University. The personality profiles of successful programmers, created by psychologists starting in the late 1960s, codified many of the stereotypes that abound today, like how “programmers dislike activities involving close personal interaction. They prefer to work with things rather than people.”

By this time, the world of computers no longer seemed to belong to the ENIAC girls. The myth of the male programmer that we know all too well was already full-fledged and seeping into the cultural consciousness. In 1968, a respected industry analyst described the programmer as “often egocentric, slightly neurotic, and he borders upon a limited schizophrenia.” (Oddly, in 1968, being neurotic made men better programmers.) “The incidence of beards, sandals, and other symptoms of rugged individualism or nonconformity are notably greater among this demographic group. Stories about programmers and their attitudes and peculiarities are legion, and do not bear repeating here,” he went on to say.

The cultural idea of a programmer went from the ENIAC girl to a beardy weirdo in sandals, and from there created a self-fulfilling prophecy and feedback loop between culture and industry that has persisted into the present day. It was helped along by the industry’s heavy reliance on aptitude tests and personality profiles, which, Ensmenger writes, encoded a functional bias against women. It wasn’t an intentional attempt to lock women out of the industry; it was merely the result of laziness. “There was widespread evidence, even in the late 1960s, that psychometric testing was inaccurate, unscientific, had been widely compromised, and was a poor predictor of future performance. Nevertheless, these methods continued to be used simply because they were convenient.”

Men may not have been out to drive women from the computing industry, but no one stopped it from happening, either. And as men began replacing women, pay in the tech industry rose. When women started joining male-dominated fields like design or biology, pay dropped. There’s a signal there: women’s work isn’t valued as highly because of gender bias. After all, if a girl can do it, “It just doesn’t look like it’s as important to the bottom line or requires as much skill,” Paula England, a sociology professor at New York University, told The New York Times. “Gender bias sneaks into those decisions.”

As Damore says, everyone has biases. And most people have good intentions. Which is why it’s so important that industries have strategies for reducing bias. When symphony orchestras started conducting blind auditions, for example, women began being hired more frequently — not out of any sort of favoritism, but because the screens prevented bias from tainting the hiring process. “Blind auditions may account for 25 percent of the increase in the percentage of orchestra musicians who are female,” write the authors of the study, economists from Harvard and Princeton.

“He frames this whole thing as if it’s about intellectual biases but he misses completely, omits, ignores, analyses of actual power and discrimination — who has been closed out systematically?” says Rebecca Jordan-Young, professor of Women’s, Gender, and Sexuality studies at Barnard College.

“They’re not new ideas”

It feels as though every few years we are doomed to rehash the same controversy, and every iteration seems to get clunkier. Murray and Herrnstein wrote a book; Larry Summers gave an unscholarly conference talk; Tim Hunt talked about the “trouble with girls” in research labs; and James Damore passed around a Google Doc full of Wikipedia links. “They’re not new ideas, there’s nothing new about trying to push back against progressive hiring and anti-discrimination policies,” Jordan-Young says.

Damore’s own attempt at this classic maneuver is shot through with a number of fallacies, the biggest of which is the structural fallacy of ignoring history while assuming that the burden of proof rests with diversity advocates to show that inequality is the result of discrimination rather than the inevitable consequence of biology. Women in the US have had the right to vote for less than a hundred years. There are some pretty good reasons why inequality persists in 2017 — big structural issues that are real and clearly linked with education, opportunity, and advancement. Why turn to bad biology when you can look at history?

And the biological differences that Damore cites aren’t even logically connected to success in computer programming. For example, he says that men are more interested in things than people, but then goes on to suggest that this makes men better suited to tech or “leadership.” But leaders don’t lead things, they lead people. And for that matter, “tech” isn’t entirely thing-oriented, either.

“Essentially, engineering is all about cooperation, collaboration, and empathy for both your colleagues and your customers,” writes Yonatan Zunger, formerly of Google. “If someone told you that engineering was a field where you could get away with not dealing with people or feelings, then I’m very sorry to tell you that you have been lied to. Solitary work is something that only happens at the most junior levels, and even then it’s only possible because someone senior to you — most likely your manager — has been putting in long hours to build up the social structures in your group that let you focus on code.”

Let’s set that aside for just a moment, however, and assume that doing “tech” means you don’t have to deal with people or feelings. Is Damore arguing that men should form a massive underclass of drone-like, thing-oriented engineers managed by a superior overclass of emotionally intelligent women? Of course not. That would be absurd.

But it’s only absurd because it’s not the way things actually are. The memo isn’t reaching for a higher truth — it is instead the expression of a reactionary instinct to preserve the status quo. Deflection: now, with graphs!

It’s tempting to believe that when we’re successful, our successes are solely because of our own merits and that everyone ends up succeeding to the extent they deserve. But meritocracy is an ideal, not a reality. “It’s something we wish was real, in our better moments we aspire to it,” Hicks says. “But it’s dangerous to think about it as a thing that operates.” It’s a simplistic view of the world that’s only comforting to the people on top.

“There’s currently very little transparency into the extent of our diversity programs which keeps it immune to criticism from those outside its ideological echo chamber,” Damore writes in his memo. He goes on to say, “We haven’t been able to measure any effect of our Unconscious Bias training and it has the potential for overcorrecting or backlash, especially if made mandatory.” Without providing any details or support or citations — only speculation and what may even be his own emotional reaction to having undergone Unconscious Bias training — Damore argues that Google’s diversity initiatives have not been adequately considered or rigorously vetted.

But the actual diversity programs at Google are fairly milquetoast. A spokesperson for Google pointed to things like a six-week increase in maternity leave (from 12 weeks to 18), active recruiting of women candidates, internships for women, and a summer camp for graduating high school seniors. These measures, and what we know about Google’s Unconscious Bias training, are all in line with what various reports and studies have recommended to address diversity in the technology industry.

Training to eliminate or mitigate bias, in particular, is key. It’s the foundation on which everything else rests. How can you address a problem if it remains invisible? To that end, the industry has poured money into numerous studies and projects to suss out what went wrong and how to fix it. Case in point: Google’s own “Women Who Choose Computer Science” study.

In that sense, Damore’s memo represents a huge step backward. It’s an arrogant and sloppy cherry-picking of assertions in order to argue against Google’s Unconscious Bias training, a program intended to remedy inequity by educating people. The memo cloaks itself in a pretense of rationality but is actually an artifact of willful, cultivated ignorance that no thinking person can take seriously.

And yet it’s a surefire bet that this exact same controversy will rear its head again in the very near future. Wherever there is inequality, people will invoke bad science to justify doing nothing about it. But with any luck, with each successive rehash, more and more people will understand this argument for what it really is: a whole lot of feelings.



Source link