LINCOLN, Neb. — The numbers of women in science and technology are dismal: Barely 18 percent of computer science degrees go to women. Women make up 11 percent of math faculty. Nearly half of the women who graduate with engineering degrees never enter the profession, or leave soon after. As the demand explodes for workers in high-tech professions who can analyze the staggering amounts of raw digital data produced every year, women barely register.
Except in one field: statistics.
The discipline, which used to have all the allure of an actuarial table, has been rebranded as part of the hot high-tech field of data science, or Big Data.
This is where the jobs are. It will take an estimated 2 million new computer scientists, mathematicians, engineers and statisticians to sort through the cacophony of data and find meaningful patterns that will help, among other things, to target customers, track diseases and find crime hot spots.
Here, women are a growing force. More than 40 percent of degrees in statistics go to women, and they make up 40 percent of the statistics department faculty poised to move into tenured positions. Several prominent female statisticians run the departments of major universities and lead major data analytics labs for industry and government. One, Susan Murphy, received a MacArthur “Genius Grant” last year.
There is no one magic-bullet reason why more women go into statistics than other Big Data STEM fields. (They are also well-represented in the health sciences.) Part of it is cultural: Research has found that women tend to be drawn to more collaborative sciences that rely on teamwork and communication.
But David Morganstein and other statisticians say they have found reasons that others in tech would do well to emulate. They include creating a welcoming environment; establishing a critical mass of more than 20 percent women, so they don’t feel like oddities or outliers; and promoting female leaders to serve as role models.
Still not satisfied, the American Statistical Association is reaching out to women to lead committees, hosting the first national Women in Statistics conference this year, and rolling out a “This is Statistics” campaign to pitch Big Data professions to middle and high school girls and minorities.
“It’s long past time that all of us in the science, technology, engineering and math fields figure out how to include more women,” said Morganstein, president of the American Statistical Association and statistician for Westat in Rockville, where most statisticians are women. “The coming need for this kind of Big Data work is so great, the supply can’t keep up with the demand. We’ve got to have all the talent we can get.”
Sexism and subtle hostility
Gertrude Cox played an instrumental role in drawing women into the field. In the 1940s, Cox became the first woman president of the American Statistical Association. (Five of the past 10 presidents have been women.) She founded and chaired the department of Experimental Statistics at North Carolina State University and helped launch the state’s high-tech Research Triangle. She is an icon to hosts of young women who felt unwelcome in math and computer science classes, many of whom have studied in the school’s Cox Hall.
Even today, beyond statistics, Big Data is largely an all-white boys' club. Google's workforce, according to its own internal audit, is 70 percent male, 61 percent white. Facebook isn't much different. The venture capital firms that fund Silicon Valley start-ups lag, too, researchers at Babson College have found: The share of women with the power to decide where to invest fell from 10 percent in 1999 to 6 percent in 2014.
Some will say that women simply prefer other paths. Others, particularly women in Big Data, say the choice to jump in or sit out this career is often influenced by sexism. A popular Web site, “Everyday Sexism in STEM,” catalogues unconscious bias and outright hostility women encounter, such as being told that women researchers are “too much trouble” to hire, or a male professor who calls the handful of women students “Miss Surname,” and male math majors who scoff at statistics as the “Woman’s PhD.”
In the North Wing of Hardin Hall, one the tallest buildings on the sprawling campus of the University of Nebraska-Lincoln, the large number of female doctoral candidates in statistics is considered a very good thing. There, women such as Marina Ptukhina, Aimee Schwab and Dola Pathak are in the majority. About half the faculty who teach them are women.
“Statistics is different,” said Ptukhina, who studied math at other universities before joining the program here. “There are so many females, you never feel like you’re alone in a man’s world. You see other women, and think, ‘Oh, they can do it, so I can, too.’ ”
Taking a break from their studies in a conference room, they talked about moving from male-dominated math courses, where they often felt isolated or put on the spot, and into statistics. “The first day, I walked into a statistics class, I noticed, ‘Oh, there are a lot of girls here,’ ” Schwab said. “But after a while, you don’t notice it. You just do your work.”
Schwab recently presented her proposal for a more accurate way to estimate disease incidents rates at an international conference in Italy.
“Other STEM departments have been under pressure to have more gender equity, but over the past 10 years, we haven’t had to lift a finger,” said Walter Stroup, a UNL statistics professor, joining the students in the conference room. “We’d get 50 applicants for a job, and 25 would be women. The diversity problem has been self-solving in our industry.”
The emphasis on teamwork and Gertrude Cox’s example inspired UNL professor Erin Blankenship to enter the field.
“Every day, you work with different people to solve a different problem,” she said. “You’re not sitting in a room by yourself proving theorems. You’re applying science to find real solutions to real problems.”
Blankenship has worked with teams on a variety of projects in environmental and social sciences, including estimating ozone exposure and avian foraging patterns. She has also won recognition for developing a statistics course for elementary school teachers — part of the effort to start early to recruit girls into the field.
Sitting in the Hardin Hall conference room, Stroup leaned back in his chair. “What could other STEM fields learn from statistics? You know, sometimes I just want to line up all the men who run these departments against a wall and say, ‘Dammit, women are competent,’ ” he said. “It’s a mind-set. They can get women into their fields if they treat them like humans and as equals.”
Fairness is just one factor
Just across the UNL campus, the heads of both the math and computer science departments, two of the most male-dominated Big Data fields, are taking a page from what has worked in statistics.
Matt Dwyer said he didn’t realize until he became chair of the computer science department last year how sharply the share of women graduates in the field had fallen — from 37 percent in 1985 to about 18 percent.
“It’s like a whole generation of young women who are not being given the opportunity to work in this very exciting field that will impact everyone’s lives tremendously over the coming decades,” he said.
He is determined to change that. His program is one of 15 chosen for the Building Recruiting and Inclusion for Diversity, or BRAID initiative. It is run by the Anita Borg Institute and Harvey Mudd College, and backed by the likes of Facebook, Microsoft, Intel and Google. The program is modeled after the small private college’s success in revamping its curriculum — making introductory computer science mandatory (and accessible and engaging) — and increasing Harvey Mudd’s share of women graduates from 10 percent in 2006 to 40 percent.
The push for diversity isn’t just about fairness, and opening these rapidly growing, stable and high-paying jobs to a broader group of people, Dwyer said. It’s also about creating better products.
“There’s a lot of good data that shows when you have a team with a diversity of experiences, a diversity of training and diversity of backgrounds, that leads you to consider a wider set of options,” he said. “And whenever you consider a wider set of options, you tend to pick better solutions.”
The UNL math department has been at work on the gender equity problem since 1989, when a dean realized that not one PhD in math had been conferred on a woman in the 1980s.
Math department leaders sought to recruit more female faculty to provide role models and, in turn, attract female students. They realized that female academics tended to be married to other academics, so they changed university policy to hire both, pioneering what’s come to be dubbed a “two body” solution.
They worked to increase women’s enrollment from 15 to 30 percent, which research has shown creates critical mass, by recruiting internationally and changing the culture. They abolished oral entrance exams, which research has shown tends to disadvantage women if all the examiners are men. They created more awards to recognize hard work, initiative and talent, and instituted open- door policies so faculty members were accessible.
And the department starts early, with an All Girls/All Math national summer camp for high school students. Every January, the university hosts one of the largest national conferences for women undergraduate math majors.
Nearly half of all their PhD teaching assistants are women, far more than at most other schools, and although women make up only about 26 percent of the permanent faculty, the figure is more than twice the national average of 11 percent.
“We created an environment in which women could succeed,” said Judy Walker, math department chair. “And if you create an environment where people are expected to succeed, they will.”
Gender gap
Women are underrepresented in science, technology, engineering and math jobs and among STEM degree holders despite making up nearly half of the U.S. workforce.
All jobs
STEM jobs
Men
Women
