Sex Difference in Reasoning Skills Is on the Decline, Research Finds

SAN FRANCISCO — Differences between men and women in mathematical and spatial reasoning ability have declined almost to zero, researchers reported here Sunday.

Marcia C. Linn, professor of education at UC Berkeley, and psychologist Janet S. Hyde of the University of Wisconsin analyzed hundreds of studies of sex differences and reasoning ability conducted during the last 20 years in adults and children and found that men’s verbal proficiency has been increasing to match that of women, while women’s ability on mathematical tests has been increasing to match that of men.

“The differences are now so small as to be negligible,” Linn told a meeting of the American Assn. for the Advancement of Science. “We should de-emphasize this issue.”

Parents and teachers are apparently not doing that, however. In a paper scheduled for later this week, psychologist Jacquelynne Eccles of the University of Colorado at Boulder will report on a 10-year study in which she found that both groups offer far more encouragement to boys to learn science and mathematics.

“Parents of boys think it is important to encourage their child’s interest in technology,” Eccles said. “For girls--unless they initiate an interest in the subject on their own--parents don’t make an effort to encourage them.”

Linn and Hyde’s study flies in the face of a growing body of experimental evidence in animals that differences in prenatal exposure to hormones are responsible for men’s enhanced aptitude at solving problems involving spatial reasoning, such as picking out a simple figure embedded in a more complex design.

Only two months ago, Canadian researchers reported that these sex differences increased and decreased on a monthly cycle corresponding to fluctuations in blood levels of the sex hormone estrogen in women. That study has since been widely criticized by women’s rights groups, who fear that its conclusions will be used to keep women out of many high-level jobs.

“I don’t know if (my results) contradict that body of evidence, but what they certainly suggest is that, in general, there is a decline in cognitive gender differences and therefore that an important part of the variance in cognitive ability is probably a result of experience or other influences on performance,” Linn said.

“But actually, I don’t think the data are very strong in that area. There have been some studies that suggest relationships between hormones and performance, but they are difficult to do and there’s an awful lot of error involved in those kinds of analyses. . . . I don’t think they are yet definitive.”

For their study, Linn and Hyde used a relatively new technique called meta-analysis that allowed them to combine data from a large number of different experiments. Doing so, they found that the differences between men and women on tests of various types of verbal and reasoning skills had typically declined by two-thirds or more since the early 1970s. In all cases except one, “the trend is toward no gender differences,” Linn said.

The one exception involves the ability to rotate objects mentally. The tests continue to show a large advantage for men in this ability, she said.

The convergence of men’s and women’s scores on tests was apparent on all tests but one, the Scholastic Aptitude Test (SAT) in mathematics. That test, which is taken by high school seniors preparing for college, continues to show a 42-point advantage for males on a scale that ranges from 400 to 800.

Interestingly, the convergence between men and women does show up on the PSAT, a form of the SAT that is given to most high school sophomores. Between 1960 and 1983, males’ advantage on the PSAT math test has shrunk from 49 points to 12 points.

Linn believes the differences on the SAT math test are caused by sex differences in self-confidence and by test questions that are biased toward males by the subject content. She noted, for example, that many of the math questions involve situations taken from sports, so that men would be more comfortable and experienced at solving them.

That self-confidence of boys is largely created by parents, according to Eccles, who studied 4,000 children and 2,000 parents in southeastern Michigan. Interviews with parents revealed that they believed science to be less important for girls and that they were more likely to encourage boys to study science and math, she said.

Parents were also more likely to exaggerate their boys’ abilities in math and science and underestimate the girls’, she said. If a son and daughter from the same family were receiving equal grades in such subjects, Eccles found, parents attributed the boy’s success to “natural talent” while the girl was perceived as “hard working.”

Families with boys were more likely to purchase computers, and boys were more likely to receive mechanical and technical toys, such as chemistry sets and building blocks, she found. In after-school and summer computer camps, boys outnumbered girls by a ratio of 8 to 1.

Teachers of science and math classes tended to reinforce the parental views, she found. The teachers paid a lot of attention to “good” boys and disciplined “bad” boys, she said, while girls were “basically ignored.”

As a consequence of these actions, by the time girls reached junior high school, they had much less confidence in their scientific ability, while boys were fairly confident in all subjects.

The parents’ implied prophecies for the children’s development thus became self-fulfilling, she concluded.