Did you read about how orchestras started to do blind auditions where the players were behind a curtain? And how the result was lots more women hired into orchestras once the biased hiring committees couldn’t act on their prejudices…women don’t have the lungs for brass instruments…women’s hands are too small for the double bass.

I did, too. It seemed plausible. Then I ran across this piece Orchestrating False Beliefs about Gender Discrimination that pointed out it was not true. I went and read the original paper Orchestrating Impartiality: The Impact of "Blind" Auditions on Female Musicians (full 56-page pdf).

Yes, the result was the opposite. Orchestras are desperate to hire more women, but they can’t put their thumb on the scale if they don’t know which candidates are the women. In fact, once auditions were behind a curtain, women were hired less. But journalists who didn’t read the full paper (journalists are generally not very good at digesting math) wrote the story they wanted anyway, probably based on the abstract. People like you and me read it. It fitted everyone's preconceptions and so, as it says in the piece:

This study has 1388 citations. It has also been featured on Freakonomics, TED talks, Reddit, Slate, New York Times, Wikipedia, and I’m sure countless other mediums. I hear about it frequently in real life when gender discrimination is discussed.

It's not entirely surprising that people were misled since the abstract of the paper says that blind auditions increase female hiring, too. I guess that's what they had to do to get it published or to get the press release their institution wanted. It's only if you look at the tables in the paper that you find that the actual data says the opposite. This is the opening sentence of the abstract:

A change in the audition procedures of symphony orchestras—adoption of "blind" auditions with a "screen" to conceal the candidate's identity from the jury—provides a test for sex-biased hiring. Using data from actual auditions, in an individual fixed-effects framework, we find that the screen increases the probability a woman will be advanced and hired.

But the results from the paper show that females are relatively more successful in non-blind auditions and relatively less successful in blind auditions. This is the exact opposite of the claim. Much of the rest of the data is not significant (in the statistical sense) and very noisy, but the authors of the paper use it to emphasize the point they want to make. They justify their conclusion by arguing that cofounding factors are more important than the actual numbers.

Women in STEM

I wrote recently about a presentation at CDNLive EMEA, 12% Is Not Enough. That is the percentage of women engineers in the UK. If you don’t look at the data, you might extrapolate and conclude that women are not doing well in general. But again, the opposite is the case. If you don’t exclude biological sciences to make the numbers come out to make a point, women do better than men in STEM in general (yes, not in engineering but they dominate biology, psychology, and some other sciences).

You may have read about women-only study lounges and women-only courses on this and that at universities. The feeling that you come away with is that women need extra help, since they must be doing badly. But women passed men in bachelors degrees years in 1982. For every 100 men who graduate in the US today, 135 women graduate. In 1987, women passed men in graduate degrees and are now at 58.7%. Women earn 111 PhDs to every 100 by men. According to the Department of Education, there have been almost 6M more female bachelor's degrees since 1982, and 13M more female degrees overall if you add in advanced degrees.

So why don’t more women do engineering? In the most equal societies in the world—Scandinavia—are the numbers much higher? Nope. They are lower. Here is an approachable article on this from The Atlantic The More Gender Equality, the Fewer Women in STEM and the chart below (from Psychological Science) regresses the data. Those two points at the top left show that under 20% of women in Sweden and Norway graduate in STEM subjects, as opposed to around 40% of Algerians, Tunisians, Emiratis (Emiratas?}, and Turks.

It seems that all you need to explain the small fraction of women in engineering is that women are so much better than men at almost all other subjects. That’s probably all you need, since if lots of women do other things, you run out of women to do engineering. This paper, Why Brilliant Girls Tend to Favor Non-STEM Careers, digs deeper into this. From there:

1. 70% more girls than boys had strong math and verbal skills
2. Boys were more than twice as likely as girls to have strong math skills but not strong verbal skills;
3. People (regardless of whether they were male or female) who had only strong math skills as students were more likely to be working in STEM fields at age 33 than were other students;
4. People (regardless of whether they were male or female) with strong math and verbal skills as students were less likely to be working in STEM fields at age 33 than were those with only strong math skills.

Or, as it that Atlantic article above says in its final paragraph:

some women will always choose to follow their passions, rather than whatever labor economists recommend. And those passions don’t always lie within science.

What About Cadence?

Orchestras have a different problem from Cadence, EDA, or semiconductor in general. More people want to play in orchestras than there are spots available. I've read that there are just a handful of orchestral violinist positions that open up every year. But conservatories of music produce many more. Psychology produces more graduates each year than there are positions in psychology—I don't mean more than the number of openings, more than the entire number of practicing psychologists. This same dynamic applies to lots of jobs where people really, really want to do them for reasons other than the salary: actor, rock musician, graphic designer, author, and so on. There are so many applicants that it is easy to put a thumb on the scale to increase the number of some group.

In tech, we have a different problem. The pans on the scale are never full, so putting our thumb on it makes no difference. We need more good engineers—we can take them all, there just are too few. So we need to persuade more people to do engineering. I've run engineering groups and our hiring has always been to try and bring anyone good on board, female or male. However, the pool is too small. It also starts young: if you don't study math seriously during your teen years you are unlikely to graduate as an engineer.

We need to persuade more people to study engineering and computer science (not just STEM in general: biologists are unlikely to design the next generation of circuit simulation). Given the different motivations of women and men, we may need two separate approaches. Luckily, AI and deep learning is making chip design "sexy" again, as you can read on pretty much any day. But I suspect we need to understand better what that word means to men and women if we are going to be successful at interesting millennials in our industry. See my post Why Millennial Engineers Should Work for Cadence.

Coding for Girls and other programs like that are great at not just teaching skills but making it easier for girls to code if they want to, as opposed to thinking that coding is not "feminine". But tech needs more programmers (and other types of engineers), period. So we may need to find more ways to get even more boys to code, too. The ultimate attraction may be when the shortage works through into salaries, and engineers get rock-star numbers. There is some sign of this happening in artificial intelligence with rumors of half-million starting salaries for deep learning PhDs. Like Wall Street salaries a decade ago, that gets that attention of people wondering what to study...female or male.

Sign up for Sunday Brunch, the weekly Breakfast Bytes email.