Much attention has been given to the gap in performance between boys and girls in mathematics skills. In a new study published in the Notices of the American Mathematical Society, Jonathan Kane, a professor of mathematical and computer science at the University of Wisconsin, Whitewater, and Janet Mertz, a professor of oncology at the University of Wisconsin, Madison, examine this gender gap and test several popular explanations. Their cross-cultural analysis seems to rule out several causal candidates, including coeducational schools, low standards of living, and innate variability among boys -- a proposal made famous in a 2005 speech by Lawrence H. Summers, who was Harvard University's president at the time. "We have pretty clear data debunking the greater male variability hypothesis," Mertz says.

What, then, is the cause of the gender gap? Like the gap itself, the cause varies, the authors conclude. Mertz and Kane, who are married, don't rule out the existence of very small biological difference, but, by comparing test scores across cultures, they indict local social factors as the likely primary culprit. Gender gaps vary from place to place, showing that cultural factors swamp biological ones.


CREDIT: Jonathan Kane
Jonathan Kane (left) and Janet Mertz (right)

Furthermore, their analysis of math-test data reveals a correlation between broader gender equity and math performance -- for girls and boys. "It seems like countries that do a good job of gender equity are also doing a good job [teaching math]," Kane says. "And we can conjecture reasons for that: Women doing better end up raising their kids better."

Doing the math

To analyze some of the theories put forth for the math gender gap, Kane and Mertz looked at internationally standardized scores for the 2003 and 2009 OECD Program for International Student Assessment math tests and the 2003 and 2007 Trends in International Mathematics and Science Study. These two datasets include data from 86 countries with a 31-country overlap. If the greater male variability hypothesis, which posits that men have a greater range of intelligence than women, is true, then that variability would persist, consistently, across all 86 countries.

Instead, "For any given country, you quite reproducibly measure the same variance ratio," Mertz says. But between countries the variance ratio changes. Persistent cultural factors, in other words, seem very important in setting variance ratios. "That was one thing that really shocked me," Mertz says.

Some scholars have speculated that coeducational schools put women at a disadvantage in learning math. But Mertz and Kane's research found that gender-segregated schools make no difference in improving math scores for girls or boys.

And while the test scores of children from the poorest countries were affected by poverty, all correlation with per capita GDP ends at $11,500. After that, gender equity -- as measured by the World Economic Forum and Social Watch -- is the only factor they studied that's positively correlated with improved test scores for girls and for boys. "It's very reproducible from exam to exam," Mertz says. "If we were willing to speculate, one thing the U.S. might do to improve math performance would be to pass the Paycheck Fairness Act and the Equal Rights Amendment to the U.S. Constitution."

Allergic to algebra

The United States ranks 31st on the World Economic Forum's Gender Gap Index and is tied for 21st on Social Watch’s Gender Equity Index. Still, the test scores of U.S. high school girls have reached parity with those of boys, and half the undergraduate math degrees awarded in this country go to women.

But after that, something goes off the rails. Just 27% of math Ph.D.s go to women. Exactly the same percentage -- 27% -- of people with careers in science, technology, mathematics, and engineering (STEM) fields are women. Women constitute a very similar number -- 30% -- of STEM college professors.

This is a problem, and not just from an equality standpoint, says math professor Rebecca Goldin, an associate professor of mathematics at George Mason University in Fairfax, Virginia, and director of research at the university's Statistical Assessment Service. "Scientific and mathematical progress relies on the best people doing their best work," she says. "If you discourage half the population [from doing science], then that part is simply not in your pool of who's the best, so the best science doesn't happen."

What's going wrong? As the study suggests, social factors are key, say people working in the trenches to improve women's representation in science and math careers. But figuring out which particular social factors make the most difference will require further study. "I think there's a whole societal issue to combat," says Deanna Haunsperger, chair of the department of mathematics at Carlton College in Northfield, Minnesota, who directs the college's summer mathematics intensive program for undergraduate women. "When girls are looking for a career, and I think girls think a lot about what their career is going to be, they're looking for a career where they can be helpful to society and make a difference. And there's just such a negative stereotype about math they don't see that as a viable option."

Data may show that girls can do math as well as boys, but the stereotype that girls aren't good at math persists and does persistent damage. Just being in a culture that believes boys are better suited to science and math than girls is enough to have a negative effect on women, as work by Joshua Aronson, associate professor of applied psychology at New York University, demonstrated. When Aronson administered a math test to "fabulously good" undergraduates in their third year of calculus -- but prefaced it with a statement that the test had never shown a gender difference -- women’s scores rocketed past men’s, suggesting that the women’s performance up to that point had been hampered by self-image and stereotype. "These are women that are as comfortable as they're going to be in their environment. They've been there for weeks and weeks, and you can nonetheless make them more comfortable by removing a stereotype, and it unleashes them," Aronson tells Science Careers in an interview.

Because stereotypes tend to create self-fulfilling prophecies, it's reasonable to surmise that knocking down those stereotypes would improve girls' (and women's) performance. One way of doing that is to raise women's economic and social standing, letting girls see smart women in high places. “I think what you want to do is elevate the position of women, so there's women everywhere doing really cutting-edge, hard stuff in the view of little girls,” Aronson says. “So they're brought up believing that women can do everything. That's certainly a different world than the one I grew up in.”

Kane and Mertz agree that attitudes in the United States need to change if we are to finish the job and close the achievement gap completely. "We live in a society where, if you met somebody at a party and said, 'I have trouble reading,' that would be a real point of concern," says Kane. "But if you tell somebody, 'Well, math was something I was never good at,' it's OK. You can even tell your kids you're not any good."

Mertz cites a shirt famously removed from racks at the American retail clothing chain Forever 21 earlier this year as an example of the unhealthy attitude towards math in U.S. culture. The shirt said, simply, "Allergic to Algebra."

"'Allergic to Algebra'?" Mertz says. "This is what's being sold in the U.S. in 2011? Whereas there's a book in Japan [for teenage girls] called Math Girls. That book is essentially an introduction to topics you would see as a hardcore math major in college, and this is a bestseller in Japan. It's in its 18th printing; they've had three sequels. Can you imagine that in the U.S.?"

Rachel Kaufman is a freelance science writer living in Washington, D.C.

Rachel Kaufman is a freelance writer living in Washington, D.C.
10.1126/science.caredit.a1100139