September 17, 2001

Are women as good at math as men?

According to Diane Quinn, an assistant professor of psychology, the mere fact that many people think the answer is "no" may itself account for the gender disparity on standardized math tests.

To examine that theory, she and a colleague conducted two studies.

"There aren't genetic differences, so we wanted to know what's going on," says Quinn, who joined the University two years ago.

What she and her fellow researcher discovered is that living in a society in which there is a "girls don't do well at math" stereotype may be enough to affect math performance. Being in a situation in which the stereotype could be applied may lead to extra anxiety and decreased performance for girls in comparison to boys.

In a recent study published in the Journal of Social Issues, Quinn and her colleague, Steven J. Spencer of the University of Waterloo in Ontario, Canada, found that if they could make the stereotype disappear, the gender difference went with it.

Some 54 male and 54 female college students were each given a multiple-choice math test. Half the students received an exam with word problems; the other half got the same material, but it had been converted to numerical equations.

According to the researchers, all the students needed the same mathematical skills to solve the problems. However, the word problems required the participants to take the extra step of forming a strategy to convert the problems into numerical equations.

Quinn and Spencer found that women and men scored the same when the problems were expressed only in numbers. But on the word problems, the men scored better.

"This study demonstrates that women have the mathematical skills and knowledge necessary to solve the problems," the authors wrote. "Something interfered, however, with the women's ability to strategize and convert the problems when they completed the word problem test."

According to the researchers, the high level of "stereotype threat," which occurs when a person is in a situation where a negative stereotype could be used to judge their behavior, impaired the performance of the women who took the word problem test.

"Because the word problem test is less familiar and there are many more steps to go through before being able to solve it - specifically moving from the words to trying to set up a solvable problem - we assume that there is more room for doubt and difficulty with the word problems," Quinn says.

In a second study, the researchers manipulated stereotype threat in order to determine whether that was the variable interfering with women's mathematic abilities. In that study, 36 college students who scored between 650 and 700 on the math portion of the SAT were given a test of multiple-choice word problems. Half the students were told that prior use of the test had shown that men and women did equally well on the problems. The others weren't given any instruction about gender stereotypes. All the participants were asked to think out loud while solving the problem.

The researchers found that when participants weren't given any instructions about gender stereotypes, women underperformed when compared to men, and were less likely to be able to formulate strategy. When participants were told that the test was gender-neutral, men and women performed equally well and didn't differ in their ability to formulate and use strategies.

"We conclude from these two studies that the knowledge of cultural stereotypes changes the testing situation for women, such that their performance is depressed," Quinn and Spencer write.

The recent journal article is the latest in a growing body of research Quinn and her colleagues have done about stereotype threat.

"It adds to studies we've already done that say stereotype threat adds extra pressure," Quinn says.

Both Quinn and Spencer are continuing their research in this area. Quinn and a UConn graduate student have recently begun research concerning women's performance in the classroom. A well established pattern shows that in high school, women and men do the same in their math classes, and the two genders choose math as a college major at the same rate, Quinn says. However, the number of women decreases in higher-level math classes. And there is a large gender gap in such math-related fields as engineering.

"Having a cultural stereotype might make women want to leave and go into other areas," Quinn says. "We're trying to determine how stereotype threat is affecting performance and whether you feel comfortable in certain majors and professions."

Quinn's research has stirred interest outside the community of psychology researchers. The Educational Testing Service, which administers the SAT and other standardized tests, did its own research about stereotypes and performance in response to work done by Quinn and her colleagues. Quinn hopes that educators and others who care about women's performance might also pay attention to her research.

"If parents and teachers became more aware of the many subtle ways they may shape math situations for girls and women, then we believe even greater changes could be made in women's attitudes towards math and their math performance," Quinn and Spencer write. "Indeed, if girls and women encounter fewer situations in which they experience stereotype threat, their increasing performance may one day break the ugly cycle of the stereotype leading to poor performance and the poor performance in turn feeding the stereotype."

Allison Thompson