POSTED 24 MARCH 2005
Studies find subtle bias
We've just discussed bias that appears when we judge the work of others. Can subtle bias affect how we see our own work? Yes, says Shelley Correll, assistant professor of sociology at Cornell University. Correll says she focuses on math because it can "cause women to disengage from the particular career path" that leads to success in science. Many math courses must be taken in order, she says, and "If a senior has not taken the correct sequence, they don't have any realistic way of getting back on the path."
To understand why many women drop out of math -- and thus science -- Correll looked at data on 18,000 high-school students collected by the National Center for Educational Statistics, including answers to such questions as: "I have always been good in math," "Math is one of my best subjects," "I get good marks in math."
Curiously, students with equal math grades and test scores didn't consider themselves equally good, and gender played a key role: Women were more self-critical than men. "We find that if you compare boys and girls, or men and women, with the same grades in math classes, and the exact same scores on standardized math tests, boys think they are better than girls," says Correll. "And since boys, on average, have higher perceptions of their math ability, they are on average more likely to enroll in calculus," a key entry point to the heavy-duty math required for most advanced science classes.
This matters, she says. "When people are taking advanced math or not, or deciding which major to choose, research shows that people draw very heavily on their own assessments... the extent to which you think you are good is very motivational" (see "Constraints into ... and Gender and ..." in the bibliography).
The consistent discrepancy between women and men on the math Scholastic Aptitude Test scores is often cited as evidence that women are inferior in science. On average, men score 537, and women 501, on that test.
But the SAT scores reflect who takes the test, says Bowen of Cleveland State University. "A lot of boys who are at-risk either don't have enough financial resources to go to college, or are not prepared or are not interested, so they don't bother to take the SAT." Since more at-risk girls take the test, that reduces the girls' average.
But there's more. "When Summers talked about clear evidence of gender differences on overall IQ, and mathematical ability, it was a mistaken statement," Bowen says. "If you look at a lot of meta-analyses on IQ or math ability, boys-versus girls, there is very little difference."
The threat of stereotypes
Test scores are influenced by what you are told before the test, according to research on "stereotype threat."
Believe sticks and stones can hurt your bones, but words can never hurt you? Stereotype threat -- hearing negative stereotypes -- says words do hurt. "If you are exposed to a negative stereotype about your group," Correll says, your test performance in cognitive areas like math and language can suffer.
To find out how this happens, Toni Schmader, an assistant professor of psychology at the University of Arizona, focused on working memory -- the brainpower you use, say, to remember a phone number long enough to dial it, while keeping in mind who you are phoning and what you need to say.
Working memory aids concentration and is essential for math.
Is your memory working?
When Schmader and colleague Michael Johns asked undergraduate "lab rats" to memorize words while doing math (see "Converging Evidence ..." in the bibliography), they found that women's working memory was impaired after hearing the stereotypes.
Schmader, Johns and Andy Martens looked at performance (see "Knowing Is Half the Battle ..." in the bibliography). They asked undergrads to solve word problems, but asked the test leader to introduce the task in three different ways:
1: The leader described the task as a problem-solving exercise. Women and men performed equally.
2: The leader presented the task as a diagnostic math test that would compare men and women. "It was a typical stereotype threat, and women underperformed men," Schmader says.
3: The leader gave the second introduction, but also explained how exposure to the stereotype might cause "increased anxiety and lead women to do more poorly," Schmader says. Women and men scored equally.
Of the three situations, then, women's performance only suffered when they believed they were being compared to men, which activated the stereotype that men are better in math. But when that situation was altered with information on stereotype threat, the discrepancy disappeared. "This is the exact pattern of performance that would be predicted by stereotype threat," the authors wrote. "If you are a woman," Schmader told us, "and find yourself anxious while taking a math test, it may be due to stereotype threat and not to your ability."
To Schmader, the research results doubly condemn Harvard President Lawrence Summers's comments: Instead of explaining stereotype threat, he exacerbated it. "I see his comments as creating a sense of stereotype threat in women who might be studying science or math," she says.
What about subtle discrimination?
Megan Anderson, project assistant; Terry Devitt, editor; S.V. Medaris, designer/illustrator; David Tenenbaum, feature writer; Amy Toburen, content development executive