By now, most people have heard about the infamous “anti-diversity” memo written by a male Google engineer (now fired) that went viral over last weekend.
There’s no point in rehashing the drivel about “biological differences” he wrapped up in scientific-sounding jargon (feel free read it for yourself here).
But it’s worth correcting the record because, for someone who’s presumably a data-driven person, he got the “facts” about women’s technical acumen really, really wrong.
Here’s what you need to know:
- Women are as competent as men in math and science, suggesting there is no native difference between the sexes. High school girls were overrepresented in advanced technical classes – 55% of students in advanced AP/Honors math classes and 56% in advanced AP/Honors science classes were girls. Research by UW-Madison and University of California, Berkeley sorted through standardized tests for 7 million students under No Child Left Behind, and found no differences in math performance between the sexes. Another study looked at three million students in a 20-year period and, once again, no significant difference between girls and boys.
- In fact, some research indicates girls have an edge in academic performance, generally. In 2016, among SAT-takers, the top 10% of graduating classes are female, and 60% of students graduating with an A+ grade point average were female.
- Scientific evidence actually suggests that there aren’t meaningful differences. One researcher looked at nearly 50 meta-analyses on differences between the sexes – in cognition, communication style, social and personality variables, moral reasoning, etc. 50% of the studies showed only small differences; in another third, they were nonexistent, this report also said.
It is true that women are underrepresented in computer science and engineering, and the problem is getting worse. What the data tells us, however, is that the “pipeline problem” is largely due to cultural factors (and, again, research pinpoints these as a lack of female role models and encouragement, societal perceptions that coding is for boys, unconscious biases that permeate who gets called on in class to who gets hired or promoted).
That’s why just 20% of computer science grads in the US are women. That’s a big problem — and the reason why is where this guy really misses the mark.
Diversity isn’t just about fairness (although that certainly doesn’t hurt). It’s not a liberal social-engineering experiment run amok. Google and other companies aren’t promoting diversity just because they’re being politically correct. It’s because there’s something in it for them.
It’s about good business.
Back to that pipeline…
Anyone in Silicon Valley will tell you that the number one challenge constraining growth is the lack of qualified engineering talent. This is the reason why we import so much talent from overseas. Between tech’s explosive growth and greater restrictions on H1-B visas currently proposed, Silicon Valley could be facing an existential crisis around talent over the next five to 10 years. How do we solve it? Recruit the people who already live here, who already have the education, the technical aptitude and the smarts, to go into computer science and engineering. And figure out how to keep them around. Again, this isn’t “social engineering.” It’s just smart business.
Diversity drives better products.
Research shows that a wide variety of viewpoints, informed by different backgrounds, ideologies, experiences, etc., is crucial to good innovation. Different people see things differently, coming at problems in unique ways that help take technology from mundane to mind-blowing. Plus, it’s likely you want your product to appeal to a diverse user base — so having people who, at least in part, resemble your users can be a simple but powerful way to ensure that happens.
The manifesto’s author argues that the benefits of understanding your audience only extend to the “front-end” (e.g. the user-interface). Having managed technology teams for more than 10 years, I can assure you that the engineers who write code without understanding who is going to use it tend to make radically worse decisions than those who understand the end user.
But the data also is clear: Diverse teams perform better, with gender-diverse teams yielding returns 15% above the mean and ethnic/racial teams yielding returns 35% above.
It’s possible that right now, this engineer feels like the aggrieved party, fired for holding an unpopular opinion. Generally, we don’t reach greater understanding without greater dialogue. However, in a company like Google, where attracting and retaining the best talent is critical to success, and where the best path for growing that pipeline is in drawing talent from the most promising pools (in this case, women and underrepresented minorities), “opinions” that a significant number of your co-workers are genetically ill-suited to do their jobs are, for obvious reasons, not only offensive but value-destroying.
At the same time, firing one engineer does not make the problem go away. There are likely many people at Google and throughout the economy who likely share the author’s views. The job of managers is not only to foster an inclusive workplace, but to foster a dialogue about why it matters for the bottom line.