Sometimes it seems like everything about me is middle. I’m a middle-aged guy; I teach middle-level courses in statistics and testing at a mid-sized, middle-of-the-road university in the Midwest (my school is in the Mid-American Conference). In my graduate training in educational measurement, my grades were...well...middling. Back in high school, I took the Scholastic Aptitude Test, the S.A.T. Back then, my scores in verbal, math, and total were--you guessed it--near the median.
But now I might have a chance to change all of that. I think that I’m going to sign up to retake the S.A.T. The word from the Educational Testing Service and the College Board--the makers of the S.A.T.--is that there are some big changes in store.
Because testing is my field, I try to keep up with these changes. Changing the name of the original S.A.T. to “S.A.T. I” and “S.A.T. II” doesn’t seem to me to be that big of a deal. As long as this trend doesn’t get carried too far and the College Board starts numbering the S.A.T.'s like sporting events--for example, Super Bowl XXVIII--I think that people in education (students, parents, college-admissions personnel, and so on) will be able to make sense of the changes. Also, the acronym S.A.T. now stands for Scholastic Assessment Test as opposed to the old Scholastic Aptitude Test. Although “assessment test” seems repetitive and redundant, I think that such a change is probably harmless and innocuous.
Another one of the changes may not be so easy to understand. It has to do with how scores on the S.A.T. are reported. The biggest change is something that they call “recentering,” which will take effect beginning in April 1995. I have to admit, though, that when I first heard about the people in Princeton, N.J., doing some recentering, I though it was kind of odd. I had always thought of the Educational Testing Service people as pretty straight-up folks. When I heard about them getting into “recentering” I wondered: “Why would these people have gotten themselves mixed up in some kind of mystical, new-age, MacLainesque, channeling-for-the-gods, finding-the-inner-child stuff for anyway?”
I was a little mixed up. High school juniors won’t have to bark like dogs or invoke the power of crystals when they take the S.A.T. As far as I can tell, there won’t even be an increase in the number of questions about isosceles triangles. To my knowledge, no one at E.T.S. is going to become a spokesperson for Ramses II. Instead, it turns out that “recentering” has to do with a major change in the score scale used to report a student’s performance on the S.A.T.
I suppose that most of us who went through schools in the United States have heard of the S.A.T., and most of us are familiar with the score scale. The scale goes from 200 to 800 on each part (verbal and math), with 500 being “average.” Of course, most people who have kept an eye on American education also know that the “average” has been slipping fairly steadily over the years. A score of 500 only used to be the average. Currently, the real averages in verbal and math are about 425 and 475, respectively. What this means quite simply is that the current average performance of U.S. students is substantially substandard, compared to what average performance used to be.
These differences between 425, 475, and 500 are--technically speaking--huge. In the past, “average” performance of 500 on the verbal and math sections would combine for a total of about 1000. Now, “average” performances of about 425 and 475 squeak in at a total of about 900. Much has been made of these score declines. Numerous analyses have been undertaken to understand the reasons for the declines; sundry educational strategies have been proposed to address the declines; apologists (pro and con) have speculated about what the declines really mean in the context of broader educational reforms, and if the declines really are declines. If nothing else, the declines have bruised the national ego. It goes without saying that we would all feel a lot better about ourselves if there had been years of inclines instead of declines.
Recentering is intended to change all of this. With one stroke of psychometric savoir-faire, the E.T.S. is going to make the 425’s and 475’s turn into 500’s next spring. Poof! The current lower levels of performance will become the new center of the scale, hence the term “recentering.” Intuitively, something doesn’t seem right about changing 400’s into 500’s with just a wave of the wand. Sure, we could all stand on a box when a physician measures our height so that we’d all look six inches taller. Or we could scratch new lines in the thermometer so that everyone with a fever would appear to be healthy. Of course, the problem with these remedies is that they’re only cosmetic--despite the “enhanced” procedures, we’d still be short and sick.
To many people concerned about American education, recentering hints that, for high school students these days, the intersection of cooking and testing isn’t limited to their home-economics classes anymore.
But, personally, the news about recentering will be good for me. I figure that if I retake the S.A.T. now that it’s been recentered, I’ll be a prime beneficiary of the windfall. I don’t need to tell you what a boon that extra 100 points would be for me. I suppose that I could go to college again, but this time in style; I’ll probably be eligible for that scholarship I just missed out on the first time. And, of course, I won’t be going to the same college. No sir-eee! I think my scores should be good enough to get me into one of those fancy-schmancy East Coast schools (no more Mister Middle-of-the-Road, Midwest stuff for me!).
And here’s the part I’m really looking forward to: I think I’m going to play intercollegiate athletics. With the extra 100 points the E.T.S. is going to spot me, I won’t have to worry about nettlesome details like academic eligibility. With the little extra boost from a Kaplan or Princeton preparation course under my belt before I retake the S.A.T., there’s no telling where I could be headed.
In fairness to the E.T.S., I am sure that they have some reasons for the recentering which go beyond just cutting me a break. I have my doubts about the other reasons, though. For example, one of the reasons, according to the president of the College Board, Donald M. Stewart, is that “after recentering, students who take the S.A.T. will immediately see where they stand in relation to other students, and how their verbal and math scores relate to each other.” Hmmmm. If this is the real reason, then we have a lot more to worry about regarding high school students. Under the current S.A.T.-score scale, students can easily see how their verbal and math scores relate to each other. For example, a 425 in verbal and a 475 in math means you are better in math; if high school juniors have a tough time with this, they deserve those scores.
Further, under the current system, students immediately see how they stand in relation to other students: If a student gets a 425 in verbal and 475 in math, he or she did worse than the student who got 500 in verbal and better in math than the student who got a 450. Why is that so tough?
Perhaps what the College Board president meant was that the average of 500 is no longer the “real” current average. Nonetheless, the interpretation is straightforward: Students who score 425 in verbal and 475 in math are “average” compared with today’s students, but still quite a bit poorer than the average student in the not-too-distant past.
Another common reason that is offered for the recentering seems to center on changes in the population of students taking the S.A.T. The argument goes something like: The population of students taking the S.A.T. is more diverse, more representative, more heterogeneous, and so on. I admit that I have never really understood this argument either (remember, I did have only average scores on the S.A.T. myself). If the population of students--however comprised--is scoring lower than students in the past, why does that suggest that recentering is in order? While the composition of the group may have changed, the meaning of a 500 hasn’t--it still represents the average level of performance of students applying for college admission when the S.A.T.-score scale was established. It just so happens that the average performance of students applying for college admission today is what has changed.
What I’m most concerned about is that this recentering phenomenon is going to get lost in psychometric jargon, while the policy implications of the change will be muddied in a sea of statistical syntheses, measurement mumbo, and technical trivia. It’s hardly a technical issue. If it were, there would never be such a change. I have a lot of respect for the people who do the technical psychometric work on tests like the S.A.T., the American College Testing Assessment, or a.c.t., and others. I’ve witnessed meetings where the concern over technical accuracy in equating test scores results in arguments about why the results of one statistical analysis yielding an average of 23.12475 questions correct differs from another analysis that shows the average to be 23.12476. These people don’t do sloppy work. But with recentering, the concern is not about a few hundred-thousandths of a point, but with 80 to 100 points. This is the psychometric equivalent of straining the gnat and swallowing the camel.
As an assessment that has had some usefulness for charting student performance across the country over time, the S.A.T. used to stand as one of the few national markers of educational progress (or regress). Now, with recentering, it is possible that only the National Assessment of Educational Progress remains as an indicator of the country’s educational health.
Nonetheless, despite the fact that recentering, on its face, does not seem to make too much sense, American education will probably survive this change, too. We can probably find a way to explicate the poor performance of a student who has S.A.T. scores of 500--just as we have found ways to make sure that no one fails a grade in elementary school, no one is denied a high school diploma, and no one gets below a C in college. Reminiscent of Lake Wobegon, “All the women will be strong, all the men will be good-looking, and all the children will be above average"--even those who aren’t. A John Cannell report on recentering will probably come and go, and those antiquarians who question the public benefit of recentering will most likely be dismissed as numb-noggin Neanderthals.
Oh yes, and now that I think of it, there will be another benefit to me as a professor of educational measurement. In my testing classes, my students usually ask me if the street talk about the S.A.T. is true. For example, they ask things like, “Should you really always guess ‘C’ when you aren’t sure of the answer to a question on the S.A.T.?” (I usually say no and explain that there are about as many questions that have “C” for the correct response as the other choices.) The biggest myth they usually wonder about is whether you really get 200 points on the S.A.T. just for signing your name. In the past, I used to wax psychometric and seize the opportunity to talk about scaling, equating, and comparability of scores.
Now I think I’ll just say, “No, you get about 250.”
