The SATs Are Bunk


Make text smaller Make text larger




The genetics of the groups involved cannot have changed in the space of a few years. A hundred millennia would be more like it. What have changed are the social conditions under which people with different pigment live. And?large-sized surprise!?how they live affects how they do on tests, as well as what food and music they choose, what color disco shirts and dinette sets they prefer and what they think of salt cod and uncles. If kids are doing less well on the standard tests of academic prediction, it's because they know less, care less or are prepared less, not that they are less.


But the plausibility of a link between genes, ethnicity and smarts remains a vital American political and moral issue that requires continued exploration. It has its origin in long-standing hierarchical notions of the great chain of human being. Self-appointed arbiters of human value have routinely ranked different groups according to various categories, ranging from missionaries concerned with whether people had souls and prayed to just one god, to cultural elitists who wonder if they dug Dvorak, to straight racists who focused on whether they needed sunscreen on their white skin and shades for their blue eyes. The Nazis carried this to both logical and biological conclusion, with phenomenal industrial efficiency. And they are with us still. Every day, somewhere in the world, when even low-tech ethnic cleansers go to work in the morning with plain old hammers or machetes, they are following in the Nazi goosesteps.


In America, this great-chain-of-human-being template was applied readily to slavery. The practice of owning and relentlessly exploiting people from Africa was justified by the claim that they were inferior beings. It was easy to pick out in a crowd the individuals defined as inferior. Their pigment made the situation easier to dominate, and before and after the Civil War the consequences of this discrimination have bedeviled the community.


The modern version of this has to do with tests. There has been a consistent argument to the effect that the reason that people with dark skin don't do as well on various tests as people whose skin is blindingly white is not that they started from rural slavery, but that they are different not only in color but competence. Again, that's why the growing gap between light and dark respondents to SAT tests is so reassuring?because it means that since the genes haven't changed, it must mean the social conditions have changed. They've worsened in the case of the dark-meat people. It is hardly surprising that the sharply increased gap between rich and poor in the USA has an impact on the work children of rich and poor people have to do, which is take tests.


And what are these tests? In a book I published in l987 (still in print, thanks, world), The Manufacture of Evil: Ethics, Evolution, and the Industrial System, I argued that tests were the quintessential tools of "the psychological-industrial complex." They permitted a large, heterogeneous society to try to identify people who would work and play well with others in the economy we created. The tests were supposed to be fair and reach out and identify the essence of an applicant?their core quality, almost at the cellular level. The tests were claimed to be then, and still are claimed to be, "culture-free." Their first major use was to select people for the U.S. Army during World War II. Afterward they became a cause for a variety of reformers, who with well-meaning enthusiasm sought to reduce the impact of the accidents of birth, class, ethnicity, gender and the like, by finding a common central factor called intelligence or scholastic aptitude or call it what you will. It was to be a method of choosing an elite of competence without simply repeating the existing economic elite.


The story of this has been told by among others A.J. Strenio, myself and, most recently, Nicholas Lemann. There are interesting components of the tale. One is where the tests are created. The SATs are generated by the Educational Testing Service in Princeton (which looks like a country club, especially in its luxurious conference center). The earliest widely used IQ test was the Stanford-Binet, a product of Princeton-on-the-Coast, Palo Alto. These locations cannot be without impact. What if the tests were formulated and produced in Harlem and Watts by people who lived there, rather than Princeton and Palo Alto? It is not wholly silly to predict that scores would be different for the groups now taking these remarkably powerful instruments of social and hence economic selection.


You also have to know how the tests are created. To reduce cultural differences, ETS removes all questions that are answered with significant difference by Catholics, rural folk, Peruvians and any other definable category. The assumption is that there aren't and can't be differences of any consequence between such groups.


However, differences continue to show up because?guess what??cultural life matters. More fundamentally, in structuring the responses of males and females, ETS removes from its lists any question that distinguishes clearly between the sexes. A kind of sanitized unisex test remains. Does one reason for the persistent differences in test results between different groups reflect the fact that ethnic or cultural groups with sharply defined sex roles and stereotypes, or different ones from the white-bread norm, will have no questions to answer that tap their particular skills and experience as males and females? Does the test penalize macho males and spaghetti-strap females, who may have ample smarts but reflect relatively unusual notions of sexuality in all its forms? Nevertheless, despite ETS' article-of-faith presupposition that there are no real and basic difference between males and females, and its best efforts at managing the tests, differences stubbornly reappear?obviously because there are differences between males and females in cognition as well other areas of life.


The problem is that these tests have no real way of translating differences of response into anything other than invidious differences of better or worse. In fact, this creation of a great-chain-of-human-being picture takes exquisite and potent forms. For example, anyone who takes the SAT receives back in an official envelope from the Educational Testing Service (good words, all) located in Princeton, NJ (prestigious town), their score on the tests, which were basically graded by machine. The recipient is told their test score?say the average this year of 1018. But they are also told where they fit in the great-chain-of-human-beings. A kid who does really well will be told that she is, for example, in the 94th percentile?that is she ranked higher than 93 out of 100 other test-takers. That can make a person bound for college feel really fine, especially since the directories to colleges in the USA contain data on the average SAT scores of students already admitted to the school, so applicants can gauge their chances.


But for every kid opening her mail in the nice part of Chappaqua to receive a 94th percentile grade, another in the dicey area stares at an envelope containing news about the inevitable reverse image?she is in the sixth percentile. Remember, she is an essential part of a normal curve of distribution, albeit in not such a desirable part of it. She practically doesn't exist. She is doomed, sixth from the bottom. Not only that: She cannot attribute her terrible score to the lucid quotidian realities of her alcoholic parents, the fact that Dad became a paraplegic in an industrial accident, the fact that her school had the worst teachers in the system. No: The test is culture-free, and she is obligated to ignore these very tangible factors. She is defective at the cellular level. How does she get up off the floor? What happens to children who are in the second percentile? Does anyone follow them and know, or care?


Instead, there is bitter national contest for the few spots at privileged institutions. To coordinate and sanitize that particular and very limited struggle, the whole student world has to be graded largely by machine. And colleges proudly advertise the high SAT scores of incoming students to assure bill-paying parents they're getting their $30,000 worth.


It used to be the case that the best predictor of success in the first year of college was parental social class. It may still be so, but now the test scores are also good predictors?even though they may be surrogates for social class generally. Clearly, fair-minded institutions need to try to find relatively fair measures of competence to fit oval pegs into oval holes. No one actively sets out to create a mean and dysfunctional system. There does not seem to be an easily generated alternative to the current ruthless steeplechase, which is becoming worse by the year as the bimodality in income and power in the country put ever-greater pressure on them that has to get more.


So you can be sure it's the youngsters of Princeton and Palo Alto who have the costly private tutors, the expensive courses that are not, according to ETS, supposed to improve SAT scores?but of course do, because they game the game?and school systems that bend what they teach not necessarily to what students need as people, but to what they need to succeed during a portentous afternoon answering questions. It is an extremely dire process. And it is made grimly worse when its own power is mistranslated into psychologically violent associations between human competence and skin pigment.


Evidently, the most highly educated single demographic group in the United Kingdom is West Indians, who provide a poor market for sunscreen peddlers. Maybe it's genetic.


Make text smaller Make text larger

Comments