S.A.T. Scores: Miserable or Miraculous?
I recently read an article by the Washington Post's syndicated columnist Richard Cohen entitled "Johnny's Miserable S.A.T.'s." I don't have to tell you its contents or how Mr. Cohen feels. This diatribe differed from other education-bashing pieces so much in vogue lately, only in that it blamed the kids themselves and their parents more than the schools for the miserable state of affairs.
Mr. Cohen shares the prevailing view that S.A.T. scores are in decline and reflect the abysmal state of education. At the risk of appearing a fool, a Pollyanna, or both, I would like to declare that current S.A.T. scores are much higher than we have any reason or right to expect.
To claim that something may be O.K. with the condition of education sounds so strange, even to my ears, that I didn't believe my own thinking. Fortunately, I used to work for the S.A.T.'s producer, the Educational Testing Service, and could call on some expert buddies still there. They affirmed my reasoning.
To understand my unexpected conclusion, one needs to look at the current state of S.A.T. scores and compare that with the state of the scores in 1941. That's right, 1941.
Currently, the million-odd seniors who take the S.A.T. each year score as an average of 424 on the verbal test and 476 on the mathematics test. In 1941, the tests' developer, the College Board (E.T.S. did not exist yet), scaled the S.A.T.'s so that the average student taking the tests scored 500 on the verbal and 500 on the math, and so that all test takers scored between 200 and 800. New editions of the sat are produced for each administration of the test, but all S.A.T.'s are equated back to this 1941 version. A 500 in 1990 means the same thing as a 500 in 1941.
Is this standard-setting important? It is crucial. Let us ask who took the S.A.T. in 1941. Were these test takers a representative sample of the country as a whole? Hardly. They were just the 10,654 students who happened to show up and take the test.
And how can we characterize these 11,000? Alas, the College Board, which originated the S.A.T. in 1926, no longer has descriptions of these students. However, one of my former colleagues at E.T.S. says we can be relatively certain that these students mostly came from affluent families, resided mostly in Northeastern states and were headed mostly for the private, Ivy League, and other selective colleges in those states. They were probably mostly white males. (My friend's credentials for these conclusions are impeccable: He took the S.A.T. in 1937 and joined E.T.S. in 1949.)
How different this group is from the test takers of today! The College Board's statistical procedures insured that this elite group averaged 500 on each of the two tests. In 1990, an incredibly mixed batch of 1 million-plus poor kids, rich kids, black kids, white kids, etc., scored 424 and 476. The colleges, searching for warm bodies, even lukewarm bodies, now that the baby boom has passed, are recruiting intensely and more of them require the S.A.T. than ever before. But these non-select, grab-bag bodies huddle anxiously on Saturday mornings and score pretty damn well in comparison with the elite college-bound of 1941. I'd bloody well call it a miracle.
And, of course, it's not just the kids who have changed. Critics do not paint kind pictures of the current teaching force, or of the curriculum they seek to impart to students. Even objective observers note that the civil-rights and feminist movements, by opening up opportunities for minorities and women, siphoned off many of education's most talented minds. John Goodlad's newly published report on the condition of teacher education is grim. A recent article by Michael Apple and Susan Jungck in the American Educational Research Journal argues that, despite the calls for professionalization, teaching continues to be "deskilled."
In spite of this, in spite of drugs, gangs, family breakdown, social turmoil, television, and the rest, our students continue to look good against the cream of 1941.
Oh, you may say, a 24-point difference and a 76-point difference don't look good to you. Remember, first, that we are dealing here with a 600-point scale. A few points mean very little. Translating the points into the number of questions missed, students today are missing, on average, about two more items (out of 60) on the math test and about 10 questions (out of 90) on the verbal test. While this is unlikely to bring joy in many quarters, it hardly leads to the conclusion, recently voiced by the College Board's president, Donald Stewart, that "reading is in danger of becoming a lost art." And it certainly doesn't mean, as Richard Cohen concluded, that today's students are dumb.
Obviously, however one interprets declines of 24 and 76 points, the fact remains that the national average for the S.A.T. has fallen. This fall occurred between 1963 and 1977, when the ethnic, gender, and socioeconomic composition of test takers changed dramatically. A blue-ribbon commission attributed most of the decline between 1963 and 1970 to this change, and said the decline from 1970 to 1977 was most likely due to "distractions"--television, Vietnam, Watergate, and the like. The commission found more reasons for the decline than there were points of decline. That should have tipped us off that education was holding its own (the commission did not ask the potentially more interesting question of how the scores had remained so stable from 1941 to 1963 in spite of a 90-fold increase in test takers).
In the 13 years since the commission's report, we have continued to become a more ethnically heterogeneous nation. The percentage of high-school seniors taking the S.A.T. has increased from 33 percent in 1973 to 40 percent in 1990 (one might think that dipping this much further into the talent pool alone would lower the scores.) The percentage of minority test takers has at least doubled. A larger percentage of our population has sunk below the barbarous official poverty line ($12,750 for a family of four). In many families, both parents work (if, indeed, there are two parents) and are less involved with the school, less able to assist their children. Given the increase in adverse conditions, one might certainly expect to see sat averages falling further still. And what has, in fact, happened to the averages? The verbal score has declined by a meager three points, the math score has risen by six points.
Suppose we had a group of S.A.T. takers in 1990 identical to the original 1941 group. How would they score? We can't answer that exactly, since we do not know the exact character of the 1941 sample. But we can make approximations. We can reasonably assume that this group was all white and largely male. In 1990, whites scored 446 verbal, 491 math. About 52 percent of this group were women, who score lower than men on both tests.
If we further assume that the 1941 S.A.T. takers were themselves the scions of college graduates, the averages go up: in 1990, white students from homes where at least one parent had a bachelor's degree scored 454 verbal, 505 math. For those who came from homes where at least one parent had a graduate degree, the scores rise to 484 and 534, respectively. In sum, as you get close to the demographic characteristics of the 1941 standard, you get very close to the original 500 that they averaged--and may even surpass them.
Yes, we may have to achieve more in education in the future to compete internationally. Yes, television may turn children into aliterates. Yes, our children may not know geography or the dates of historical events (which strongly suggests they weren't taught, since kids, even our putative ignoramuses, tend to learn what they're taught). But on those skills tapped by the S.A.T., students continue to perform better than we have any right or reason to expect.
Why, then, do politicians, journalists, even some educators continue to use the S.A.T. as a club to beat up on education? Beats me. When people persist in a course of action in spite of evidence to the contrary, it suggests ulterior motives.
Obviously, there is a dark side to the data presented above: If we take it that the skills on the S.A.T. are important, and they are strongly correlated with success in college, it means that we have not brought our dominant minorities, blacks and Hispanics, into the mainstream of academic culture. (Those who would explain the differences among ethnic groups in terms of "bias" are left with the difficult job of explaining the extraordinary performance of Asians, 43 percent of whom in 1990 still said English was not their native language.)
To say that today's white students score well in comparison to an elite of 50 years ago is not to say there aren't major problems in bringing minorities into full participation. But that's quite a different focus than using scores, as many are wont to do, to paint all of education unthinkingly with the tar brush of failure.
Vol. 10, Issue 12, Page 36Published in Print: November 21, 1990, as S.A.T. Scores: Miserable or Miraculous?