What a French Middle School Reform Tells Us About How to Improve our University Admissions System
Several decades ago, French educators noted an unusual development in middle schools (collèges), mostly in well-to-do neighborhoods: record numbers of students were suddenly opting for German as their principal foreign language. Was it because of West Germany’s economic prowess? A desire for post-war reconciliation? A craze for the poetry of Goethe and Schiller? In fact, the answer was much more prosaic. The French government had recently decided to end ability tracking in middle schools in the hope of creating a more level playing field for students from disadvantaged backgrounds. But upper middle-class parents quickly realized that the schools would now group children together according to… foreign language choice. So: auf Wiedersehen, Englisch. Guten Tag, Deutsch!
I recalled this episode when thinking about the apparent demise of the SAT test, as discussed in this recent Ross Douthat column. I have no love whatsoever for the SAT. I remember the miserable mornings I spent as a high school junior and senior filling in little ovals with a number two pencil in response to what seemed often inane questions.
At the same time, though, the current arguments about the test strike me as rather hilariously misguided. Is the test biased against minority students and poor students? Well, yes. Obviously. The SAT measures students’ command of the linguistic and mathematical skills that it deems necessary to flourish in university, and, beyond university, in a wide range of relatively high-level jobs in American society. Especially in the language sections of the exam, students whose parents have themselves gone to elite universities and work in high-level jobs will already have a familiarity with these skills, giving them a leg up over their competitors.
In addition, there is the entirely obvious point that parents want to do everything possible to help their children succeed, and well-to-do parents command both far greater resources and far more know-how for this purpose than less privileged ones. They know which school districts and neighborhoods to live in. They know how to get their children into selective private schools and they have the resources to pay the sky-high tuitions. They will pay for tutors and test prep classes and college counselors, at a cost of hundreds of dollars per hour.
At Princeton, where I am lucky enough to teach, my best students have come from all sorts of backgrounds, including from poor minority families and wretchedly bad public high schools. But it is hardly a surprise that, on average, the students who show up best prepared on the first day of college come from places like Exeter and Andover, New Trier and Palo Alto High School. Wealthy parents, after all, are generally not fools when it comes to ensuring their children’s futures. If these high schools didn’t provide large advantages, the parents wouldn’t send their children there.
And as the French middle-school example suggests, no matter what system of selection that elite universities might try to devise, in pursuit of a more egalitarian admissions system, wealthy parents will soon find out how to game the system. Because that is what they do. It’s not fair. But it’s all too human.
A new story in ProPublica illustrates this point perfectly. It describes online services that are now charging families as much as $10,000 to connect high school students with mentors who will include them as co-authors on scholarly publications. Not surprisingly, the research in question is often of dubious quality, and the students’ contribution uncertain, but college admissions officers lack the time and expertise to evaluate the work. And with the SAT apparently on the way out, inclusion on a scholarly paper has become yet another way for elite families to give their little Zacharys and Emmas an advantage in the ever more brutal competition for the most selective schools. It is the “French middle school principle” at work.
Ironically, despite the fact that students from elite backgrounds score better on the SAT, that test actually helped keep the playing field level, for a very simple reason: admissions officers at selective universities knew its limits. They did not put blind faith in it as a measure of ability and preparation. Yes, public universities with limited resources often used SAT scores as cut-offs, automatically rejecting anyone who didn’t reach a certain combination of scores and grades, and automatically admitting anyone who scored above a different combination. Yet even there, grades factored into the equation as well, and for a large range of students in between the benchmarks the admissions officers exercised considerable discretion. At the most selective private universities, SAT scores guaranteed nothing, and many applicants with perfect scores ended up rejected. At all these universities, the admissions officers, knowing the limits and biases of the SAT, strove to compensate for them in deciding whom to admit. All too often, critics who excoriate the SAT as biased have forgotten this simple fact. In pushing to eliminate the SAT, they in fact helped to create a system that is becoming more opaque and less fair, not the reverse.
If we recognize that elite families will always find a way to game the system, then two points follow, if we want to create a more genuinely progressive, egalitarian admissions system. First, don’t rely on the system. Rely on well-intentioned universities. The SAT reflects the vast inequalities in American society. It didn’t cause them. Rather than trying to design a perfectly equitable test, or to eliminate all tests, hire admissions officers who recognize how inequality factors into student preparation, and who work to give less-privileged students a fair shake.
Second, given that any test will reflect the inequalities of American society, at least have a serious test. The SAT really is a very bad test. It’s tedious, often inane, and only evaluates reading, writing, and fairly basic math. Students waste enormous amounts of time and energy preparing for it. (SAT subject tests and AP tests are better, but they are optional). Why not replace the SAT with a serious test that evaluates both basic skills and also basic knowledge of history, civics, geography, economics, culture, foreign language, and science? Make it a required national exam, akin to the French Baccalaureate or the German Abitur? Of course, in our hyper-polarized country, such a test would immediately become a political football. But its design might still be an occasion to think seriously about what sorts of skills and knowledge students should have today as they start university.
Sure, the Exeter and Palo Alto kids would do much better on this exam, on average, than kids from poorly funded inner-city high schools, just as they have done much better on the SAT. But admissions officers could, hopefully, recognize the many factors that contribute to this differential and make their decisions accordingly, just as they have done with the SAT. And at least the students would be learning something serious in studying for the new test, rather than just memorizing vocabulary, figuring out how to decipher poorly written questions about banal reading passages, and reviewing junior high school math. And the new test might even provide some useful information about how well-prepared admitted students really are for college.
A more radical solution to the issue, of course, would be a more complete overhaul of American higher education. My colleague Matthew Karp recently pointed to a fascinating debate that took place during the Massachusetts Constitutional Convention of 1853. Some representatives called for the state to take over Harvard: to make it public, to vastly increase the size of the student body, and to abolish tuition. Had the motion succeeded, the story of higher education in this country might have been quite different. Knowing all too well both the strengths and weaknesses of the French system of higher education, which is close to the model these reformers proposed, I won’t say for sure that things would have been better. But whatever the case, absent a very unlikely progressive wave in upcoming national elections, this solution is unlikely in the extreme.
So design a serious test, and count on well-intentioned universities to make intelligent use of it.
I too think that eliminating the SAT would exacerbate inequalities rather than alleviate them. I like the solution of replacing the test with a more content base exam like the French Baccalaureate. In Israel students take a few subject tests at a variety of levels (based on individual preferences) plus an SAT style test, so that the admission process is less "subjective." (Of course, affluent families trump this system as well). But in any alternative we come up with we need to account for cost. Sure, it would be great if every application will be reviewed by an intelligent, considerate, and fair minded person who is not exhausted by reading 10 similar applications in the previous few hours. That would be possible at elite institutions, perhaps. They have the funds to pay all these salaries. But most institutions will move to subcontract the application process to AI which will make impersonal decisions about admissions based on some parameters fed into the program. And once again, as David writes, privileged parents will find the right formula to trump the AI. The subject centered solution, however flawed (and it is very flawed), offers the best path.
It's hard to argue with your reasoning here, David. But I think it's missing an analysis of the pressures on admissions officers and the unfortunate truth that not all universities are entirely well-intentioned. Or perhaps it's more accurate to say their good intentions do not conform to my good intentions.
You are right that SATs do not cause but reflect the savage inequalities of American life. In a better world they (or preferably yet the better test you would design) would simply be a measure. They might point to outliers that would allow admissions officers to take notice of people they might otherwise have missed. To take risks and create opportunities. They would help make the system more equal rather than more unequal.
The problem as I see it is that the admissions officers do not have a very free hand. They experience relentless pressure from their bosses to increase average SAT scores so that the institution's ranking can go up. I would guess this kind of pressure is relatively low at Princeton. (Princeton and a handful of truly elite places after all measure the quality of the ranking not the reverse. Any ranking that would put them in the mid-tier would be ignored.) Alas, the pressure is quite a bit higher at a place like JHU (where its administrators want to be Princeton or maybe just at Princeton) and it's exponentially worse at places falling further down the US News rankings. And given the weight US News puts on SAT scores that means universities are doing just as much gaming as applicants and their parents.
All that leads to perverse situations. Many amazing kids who would flourish in a first-rate college or university are not able to get in because their collective SAT scores would bring the institution's rankings down. Sorry wish we could help. Admissions folks can make one or two exceptions, maybe, but not many. Better to play it safe, and safe means higher SAT scores and since that is so completely correlated with class and race and other factors it means exclusion.
Even worse, as I'm sure you know, many places will give scholarships to attract the high SAT scorers coming from families that don't have the highest need but will happily take the money. Parents of those kids can shop around as though buying a used car. All this diverts scarce financial aid to artificially boost SAT rankings by favoring families that, again, correlate to high SAT scores.
Anyway you get it. I don't see how creating a better test would solve this part of the puzzle. The challenge here is getting universities to abandon the moronic rankings systems that rely on flawed metrics. But sadly these rankings systems seem to be how senior administrators prove their value to boards of trustees that need metrics because that is the language they speak. And given that dynamic, I fear we are stuck with this system.
All that is why I think it makes sense to give up SAT scores. It's true that we then get a more opaque system, and maybe the cure will be worse than the disease. I guess that remains to be seen. But surely seems to mean that we have become even more dependent on the good intentions of the admissions folks.
Always love reading your stuff, mon ami!