Using large national data sets, investigators at the University of North Carolina at Chapel Hill have discovered that -- contrary to previous research -- U.S. boys hardly surpass U.S. girls in mathematical ability.
"Whether boys' mathematical skills are superior to girls' has been a controversial topic among social scientists for decades," said Erin Leahey, who is completing a doctorate in sociology at UNC. "Strong evidence for a male advantage comes from the Study of Mathematically Precocious Youth, which began in 1972."
In that research, gifted seventh-graders were invited to take the mathematics section of the Scholastic Aptitude Test (SAT), which was considered a test of reasoning skill for the children since it involved material students had yet to cover in their coursework, Leahey said. The resulting gender differences were dramatic.
About four times as many boys scored above 600 as girls did, and about 13 times as many boys scored above 700, she said. A male advantage was also evident among college-bound high school seniors. Every year between 1966 and 1997 males scored higher than females on the math section of the SAT.
In their new study, Leahey and Dr. Guang Guo, associate professor of sociology and a fellow at the Carolina Population Center, examined for the first-time gender differences in test scores from elementary school through high school among U.S. students. They also analyzed sub-samples of high-scoring students and different areas of math, such as reasoning and geometry.
Among the findings were that girls had higher average math scores than boys until about age 11 and higher reasoning scores at ages 11 to 13.
"Despite relatively equal starting points in elementary school, we found that boys showed a somewhat faster rate of acceleration in math," she said. "By the 12th grade, this resulted in a slight gender difference, which is more pronounced in geometry."
The bottom line is that the UNC work does not support earlier research showing that large gender differences emerge by junior high school or exist at all, Leahey said.
"The largest gender difference between boys and girls was late in high school and that was only 1.5 percent, which is very small," she said. "Based on conclusions from the earlier work, we had expected to see a gap by junior high school, but we didn't see any difference at all then."
A report on the findings appears in the current issue of Social Forces, a sociology journal.
One major advantage of the new investigation was that it tapped both the broadly representative National Longitudinal Study of Youth and the National Educational Longitudinal Study, Leahey said. Most previous work on the subject relied on small, selective data sets. Another strength was that she and Guo covered such a long time span in math skills development.
"We also broke down the scores to see how students did specifically in algebra, geometry and reasoning rather than just relying on the overall math scores," Leahey said. "We found out how students did on every question. To our knowledge, that's never been done before."
At first the two thought that perhaps they were doing something wrong because they couldn't find the large differences other studies reported.
"After uncovering hardly any differences in general math, we explored alternative measures, subsets of students and modeling techniques," she said. "Still we never found the large male advantage we expected."
The researchers planned and carried out their three-year study because they felt earlier scholars might have made unwarranted generalizations about math differences based on samples that were too limited, she said. Such research is important because perceptions of girls' lower ability in math can strongly affect career choices and advancement. Almost all top mathematicians are male.
"We didn't refute the earlier studies entirely, but we think we have shown that previously reported gaps are exaggerated and that people need to be more careful when making broad generalizations on this important issue," Leahey said.
UNC's Carolina Population Center and the W.T. Grant Foundation supported the new study.
Note: Leahey can be reached at (919) 843-6876 (w). E-mail: email@example.com Contact: David Williamson, (919) 962-8596
By DAVID WILLIAMSON
UNC News Services