THE SCHOLASTIC Aptitude Test (SAT), which over the years has been frequently accused of one kind of unfairness or another, is now under attack from critics who say the test is biased against women. Males score better on the average than females in both the verbal and mathematical portions of the test, which is required for admission at many colleges. In New York State, most recently, two consumer groups -- FairTest and the New York Public Interest Research Group -- have pushed a bill through the state legislature establishing a commission to investigate the test for bias, after studies had shown that this score difference translated into thousands of dollars in lost scholarship money for young women. (New York State awards its Regents scholarships and the lucrative Empire State Scholarships solely on the basis of SAT scores, a practice discouraged by the test sponsors.) Systematic differences in SAT score levels between groups do exist -- boys on average outscore girls, whites do better than blacks, well-off students better than poorer ones -- and the differences are cause for concern. But the bias argument is fraught with illogic.
Supporters of the New York legislation have had some trouble explaining even in the most hypothetical terms how a test question -- a math problem, say -- could be "biased" against female test-takers. Their attempts lead to absurdities, such as the suggestion that math word problems on certain subjects -- sports, say -- are somehow more difficult for females than males. Such arguments threaten to become circular by implying that there in fact is a capability difference between girls and boys -- which, if it were true, would make it impossible to argue unfairness from a score difference. Reformers likewise tend to blame score gaps between racial groups as manifesting some subtle racial bias in the questions themselves.
A far more likely explanation, though, is that the SAT reflects discrepancies in education. Children from lower-income families and neighborhoods on average get lower-quality preparation than those from more affluent backgrounds. In the case of women, differences in who takes the SAT and when -- for instance, girls typically take it after fewer years of math courses than boys -- may go a long way toward explaining the difference. Dismissing these possibilities, seeking to change the tests in some unspecified way to mask such gaps, not only constitutes shooting the bearer of bad news but would eliminate an essential source of information on the problem.
The reminder that the SAT does reflect factors other than raw ability (preparation, socioeconomic status) is a useful corrective to the temptation to use test scores as absolute cutoff levels for college admission (rare nowadays) or scholarships (as in New York). But it should be remembered that most of the alternatives to SAT scores -- such as scores on Achievement Tests in specific subjects -- are even more sensitive to these same outside factors, and admissions tests based upon them may lead to less diversity, not more.