Three years ago, a shudder went through the hearts of female students and their tuition-paying parents when research by a women's advocacy group concluded that the Educational Testing Service's Scholastic Aptitude Test, or SAT, is biased against girls and women.
Each year, more than 1.7 million students who want to go to college, many of them seeking scholarships, plunk down $17 to take the test. Fifty-two percent are female.
A riveting study of SAT scores for verbal and mathematical skills in 1989 by Phyllis F. Rosser, director of the Equality and Testing Project in Holmdel, N.J., showed that the gap between SAT scores of females and males was 57 points, with girls averaging 878 points out of a possible 1,600 and boys averaging 935.
"Girls aren't dumb, so I question the test, which does not accurately predict the work female students can do in their first year of college," said Ms. Rosser, whose research was financed by a grant from the Women's Educational Equity Act of the U.S. Department of Education. "To me, that's consumer fraud."
But in the intervening years, SAT tests have not changed. According to the Educational Testing Service in Princeton, N.J., more than 1,800 colleges use the SAT, a slight increase in the last decade.
"My feeling is that the test is as free of gender bias as we can make it," said Nancy W. Burton, director of research and development at the testing service. "There is a lot of gender bias in the U.S. culture and in our language, so it's a continuing battle to try to find sources of subtle bias and get rid of them. We do the best we can."
But Ms. Rosser is back, this time with a new study for the Center for Women Policy Studies in Washington. In 1989, the center published her study, "The SAT Gender Gap: Identifying the Causes."
Ms. Rosser said her latest research confirms her earlier findings.
"Unless the questions that create a large gender difference are removed, SATs are a terrible way to determine who's going to college," she said. "High school grades are the best predictor. Why put people through the agony of this biased test?"
Another blow against the SAT is delivered by FairTest, a Cambridge, Mass., advocacy group that charges that boys win more National Merit Scholarships because of the test that's used, the Preliminary Scholastic Aptitude Test, or PSAT.
"The bias against women and minorities in the tests is so deeply embedded you can't get it out," said Sarah Stockwell, FairTest university test coordinator.
"It's a glaring example of its misuse. Girls lose scholarships, are not admitted to top-choice colleges and don't get into special programs. That's why some of the other large scholarship programs use high school grade-point averages.
"Minorities are hit harder than women as a group: There is a 60-point differential in test scores between women and men, and 200 between blacks and whites."
SAT's Ms. Burton replies that there is "no proof" that women and minority groups are at a disadvantage in multiple-choice tests. And to statements that SATs are not needed, she responds, "Taken together with high school grades, the SAT helps provide accurate information to predict college success. Separately, neither one of them does."
But another study -- this one of 12,000 college students by researchers Howard Wainer and Linda Steinberg of the Educational Testing Service itself -- shows that males with B's in college calculus scored an average of 37 points higher on the SAT than female classmates with the same grade.
Mr. Wainer and Ms. Steinberg conclude there is "evidence of differential validity by sex on the SAT . . . women score lower on average than men of comparable academic performance."
Leslie R. Wolfe, executive director of the Center for Women Policy Studies, finds it "absolutely astounding the Educational Testing Service continues to ignore this research, even by its own people, which finds there is bias in its test. The real issue is that the SAT has one purpose -- to predict college students' first year -- and the SAT doesn't do it."
Ms. Wolfe, who has a doctorate from the University of Florida at Gainesville, has some advice for the testing service:
"Given that the service has an incredibly talented group of researchers and test-makers and prides itself on screening every question before it actually is used, we believe it could very easily eliminate any question that produces this big of a difference in the answers of females and males."
But the testing service's Ms. Burton says the SAT is not the problem.
"There's growing evidence that there are serious gender differences in the way kids are treated from infancy through school, and those differences have an impact on what young women and men learn," she said.
"It makes me crazy to be blaming the test when what we need to be doing is to take care of the underlying inequities in the way boys and girls are treated in school and at home."