^{The Bottom Line} If I Ran Admissions at an Elite School

Editor’s Note: The following is an excerpt from the essay, “Making Admissions to Elite Colleges Fair.” The full essay can be accessed here.

If by some quirk of fate I should have complete control over how admissions at an elite school get decided, what would I do? It’s the old hypothetical of if I were king or dictator or grand poobah. Hypotheticals like this often have no traction with reality. But bear with my proposal for running admissions at an elite school. Once the proposal is on the table, we can turn to its practicability — whether it stands any chance of actually being implemented. 

My goal in changing admissions at elite schools is to make them fair. Right now they are unfair, privileging students who’ve had every opportunity to build their knowledge, skills, and accomplishments. The point of my proposal, then, is to level the playing field so that students who have not had those privileges, but are able to hold their own, are allowed on the playing field, which is to say, be admitted to the elite school in question. There’s no perfect solution to making admissions to elite schools fair — the privileged always have advantages, which is what makes them privileged. But there’s much that can be done to redress the present obvious unfairness. Entrance exams can play a pivotal role here, especially when combined with other evaluative measures. 

Let’s be clear, though, that we don’t want to admit students that can’t do the work. This has been the problem with DEI and affirmative action, namely, that it admits students who cannot handle the demands placed on them in their course of study at an elite school. This always ends badly, and it ends in one of two bad ways. On the one hand, schools may maintain their high standards. In that case, students admitted who cannot do the work flounder. They then experience what is called relative deprivation, which is to say that even though they could do well at a less demanding school, they feel bad about themselves because they cannot keep up at the more demanding elite school. 

On the other hand, schools may lower their high standards. Thus, to buttress DEI and affirmative action, they reward students for reasons of identity (ethnicity, race, class) rather than merit. These students, who would otherwise flounder, thus succeed (or more accurately, get by) because they get to take undemanding classes — and what exactly is the value of such success? Princeton has recently taken this route by no longer requiring classics majors to know Latin or Greek. Harvard is now teaching remedial math — high school math that Harvard students in the past would be expected to have mastered on entry. MIT, I’m told, has also lowered its standards. The mother of one of my son’s Caltech baseball teammates was an MIT grad. She described efforts by MIT alums like herself to restore the standards back to where they had been. 

If, then, I were head of admissions at an elite school, what would I do? I would construct an admissions test that would be the primary tool for admitting students. I would make it a six-hour exam, three hours in the morning, three in the afternoon. It would need to be given in a secure location where students would have no chance of cheating by accessing external sources of information (via smart phones, crib notes, etc.). In fact, given the pervasiveness of AI these days, a crucial element of this exam would be to test how well students can think on their own feet, which is to say without any props, especially from AI. Indeed, students who have made a habit of letting AI do their work can expect to perform badly on this exam. 

I’m imagining students would show up at a secure test-taking facility where they would work on dedicated laptops (stripped down machines without internet access). If they need calculators, note pads, or other materials to help with taking the test, these would be provided. Test takers would need to be thoroughly checked over, even more thoroughly than by the TSA at an airport, for any evidence of surreptitious information sources, such as earpieces or smart glasses. These, as well as more obvious sources of information, such as smart phones, would need to be absent from the test-taking environment. Those who would slip them in surreptitiously would, for attempting to cheat, be excluded from taking the test.

Administering the test could pose logistical challenges, especially if there is only one day of the year when it is given. But technology makes it doable. Also, it’s unlikely that people will take the test indiscriminately. Students who set aside six hours on some given Saturday to take an admissions exam for a particular elite school are serious about trying to get in; conversely, those who are not serious won’t even bother trying to take the exam. Still, I would set a minimal threshold for taking the test, such as a GPA of at least 3.5 and an ACT score of at least 30. I might even go with a GPA as high as 3.8 and an ACT as high as 32. But I don’t think I’d go higher. 

I would want to look at unweighted GPAs rather than weighted GPAs as a way to limit grade inflation. I would also want to set a GPA of 3.5 or 3.8 as a floor only for the junior and senior years of high school. Such GPAs are way lower than the cut-off for most elite schools, but the point of this “low bar” would be not to overvalue grades. Also, it would discourage obsessive concern about grades among students applying to elite schools — there’s more to life than holding a super high GPA.

Some very bright students only get their acts together as juniors and seniors in high school. I would want to see a positive trendline with grades in the junior and senior years, not putting undue weight on grades during freshman and sophomore years. I would also want to see some diligence on the part of applicants so that it’s clear that they were at least trying in high school (if they blow off high school, they’ll probably blow off college, which as head of admissions I wouldn’t want to encourage). That said, I know of brilliant people who don’t do well in high school because they are so bored with the material (one such person is my friend Chris Langan, who in the late 1990s scored the highest of any American on IQ tests). 

I also would not want to admit only those students with super high ACTs or SATs. In my view, someone who can take the ACT and score a 32 with minimal preparation for this test is comparable to someone who scores a 35 by going through an intensive test-prep course. Note that a score of 36 with the ACT is highest, and scores ranging between 33 and 35 are typical for elite schools. In any case, I would have GPA and ACT provide minimal quality assurance, but then put the main burden for admissions on the admissions test.

The Proposed Admissions Test

The admissions test I’m proposing would break into two three-hour portions. The first three hours would be standardized in the sense that each test-taker would attempt the same set of questions. The second three hours would be interactive, in which test-takers would be asked probing questions and then be able to demonstrate deeper levels of knowledge depending on their answers. For instance, a question might ask, “Do you know anything about ordinary differential equations (ODEs)?” If the answer is yes, further questions would be posed to see just how deep this knowledge of ODEs goes. If the answer is no, no further questions on this topic would be asked.

The precise details of what would be on such a test would need to be worked out in consultation with faculty, who will be teaching the students taking the test and will want the test to identify students best able to excel in their classes. The actual test would be refined over time as high test results come increasingly to correlate with student thriving at the elite school in question. Certainly, the test should avoid asking students questions that shift the focus from academic abilities to autobiographical descriptions that, depending on cultural predilections, will put some students in a more favorable light than others. In other words, the test should give no comfort to DEI, CRT, affirmative action, etc.

Although I can see such a bespoke admissions test for an elite school taking a number of forms and asking different types of questions, if it were up to me, I’d be sure to require one or more writing samples, created on the fly by test-takers on topics of general interest. E.g., “Should college be free?” “How do talent and perseverance relate to success?” “By what guidelines should social media companies moderate content?” All the writing samples currently in a typical college applications packet are suspect because they can be written and edited by people other than the actual applicant. The only exception was the SAT Essay, but the College Board discontinued it in 2021. The writing sample(s) in the admissions test I’m proposing would therefore provide a real test of writing ability.

Other things on this admissions test that I would like to see there would include the following:

• The test would include a deep dive into vocabulary, grammar, and style to determine just how extensive (or not) a student’s knowledge of the English language is. Free Rice, which gives rice to poor nations based on online users taking and scoring correct answers on various tests, offers a vocabulary test similar to the one I have in mind (go here). Grammar and style could be tested by putting the test-taker in the role of an editor and requiring passages with substandard English to be improved.
• The test would include sequence completions, analogies, logical reasoning, and puzzles. The point is to see how well students can draw connections based on common knowledge. The world’s so-called shortest IQ test is in this vein. The Miller Analogies Test was also in this vein, administered back in the day to prospective graduate students. But like so many tests used for admissions, this test was also discontinued around the time of Covid (in 2023). Yet in discontinuing this and other similar tests, elite schools weaken their ability to admit students fairly, looking instead to factors that students, especially privileged students, can all too readily manipulate. 
• The admissions test should gauge knowledge in a variety of areas that are important to both academic and practical life. It should test knowledge of basic facts about Western Civilization — hat tip here to E. D. Hirsch Jr.’s shared foundation of factual knowledge, which he calls “cultural literacy” (compare Natalie Wexler’s The Knowledge Gap). Successful test-takers would need to demonstrate a solid knowledge of American civics. As for practical knowledge to be tested, this could include money management, etiquette, gun safety, self-defense, everyday law, etc. etc. — the possibilities are wide and varied. Applicants with such practical knowledge should, in my view, be rewarded for that knowledge. 

The first three hours of the test would be standard, with everyone answering the same questions. In its grading, one test score would always be comparable with another, the higher being better. Based only on this half of the test, the highest scoring students would be admitted, and so there would be a cutoff below which students with certain scores would be refused admission. This half of the test would be designed to make fine-grained distinctions among already high-performing students, a distinction that gets flattened with the ACT and SAT tests. That’s because these standardized tests can’t meaningfully distinguish among students once their scores place them in the 98th or 99th percentile. 

The second three hours of the test would, by contrast, be designed to interact with students, probing their knowledge. A student could be asked about ability in foreign languages. “Do you know Mandarin?” “Yes.” “Okay, then what does the Mandarin word “谢谢” (xièxie) mean in English?” (It means “thank you.”) The 1869 Harvard admissions exam probed knowledge of Latin and Greek, but with digital technology it is possible to determine proficiency in languages much more quickly and efficiently (cf. the five levels of proficiency as measured by the American Council on the Teaching of Foreign Languages, or ACTFL). 

The interactive portion of the test is meant to mirror human interactions with experts. Direct human interactions can very quickly determine knowledge and competency. Unlike several-day batteries of tests as needed to pass the legal bar or the medical boards, oral exams allow for very quick and reliable determinations of knowledge. When as a math grad student I had to pass two foreign language exams, I chose French and German. With the French exam, I was required to translate a passage from a French mathematics text. I passed but it took time to complete the translation because my reading knowledge of French wasn’t great. With the German exam, I showed up to the German examiner (mathematician Israel Herstein) and mentioned that I had lived in Germany. He said, “Okay, start speaking in German.” I uttered a few sentences in German. He said, “Okay, you pass.” AI can now substitute for such oral examiners. That’s how I would use it on the second half of the admissions test. 

Note that I don’t foresee grading such tests as difficult. Many of the questions, certainly on the first half, could be fill-in-the-blank or multiple choice, which are easy to grade. On the second half, I could see AI/LLMs doing the probing and the grading, though there would need to be quality assurance to prevent AI “hallucinations,” as they’ve come to be called. Some human graders would be needed, such as for the writing samples. But it should be possible to automate most of the grading of the admissions test I am proposing.

Using This Test in Admissions

Deciding admissions to an elite school in light of the test I’ve just outlined would not be like using a cookie cutter. Students taking the test would need to do well enough on the first half of it to ensure that they can “do the work,” being able to take the required courses and do well on them without excessive struggle or hardship. But once this portion of the admissions test has narrowed the field to qualified applicants, the admissions committee will still need to do some winnowing. True, students willing to take a six-hour bespoke admissions test to get into an elite school will be far fewer than students who apply only to schools where it’s enough simply to send digital application packets from the students’ hard drives. So, even if there is winnowing, it’s likely to be far less than the winnowing that most admissions committees at elite schools have to do now. 

The second half of the exam will facilitate this winnowing, helping qualified students to be admitted who exhibit a true diversity of interests and abilities. Presumably, it would enrich a campus to have students who have mastered different languages besides English. Their mastery of these foreign languages could be tested and rewarded with admissions. Likewise for students adept at so-called “dead languages,” such as Latin and ancient Greek. Deep knowledge of various STEM fields could likewise be tested and rewarded. An extensive knowledge of literary classics (as gauged by knowledge of characters, plot lines, and the ability to associate quotes with their authors) could also be tested and rewarded. 

Things important to the campus culture could also be factored into the second half of admissions test. Caltech, for instance, has a pranking tradition, as I learned when my son applied to the school and showed me a question on the application where he was asked to describe notable pranks in which he had taken part. Caltech is famous for its clever and elaborate pranks. The Great Rose Bowl Hoax of 1961 stands out, which even has its own Wikipedia entry. During the 1961 Rose Bowl game between the University of Washington and the University of Minnesota, Caltech students hacked into the stadium’s flip-card section that displayed choreographed images held up by Washington fans. Caltech pranksters reprogrammed the cards to spell out “CALTECH” in huge letters:

Nothing about the admissions test I’ve outlined would prevent elite schools from admitting students with special skills that they want to see represented on campus. If a school needs a student to play a bagpipe, then it needs to admit a bagpiper. But such a student would need to be a competent bagpiper and also have done sufficiently well on the admissions test to prove capable of the coursework. Sports also fall under such special skills. If the elite school in question needs to field a sports team, then it needs to admit competent athletes in the sport, but as always with the proviso that these student athletes can “do the work.”

In the admissions process, I find it helpful to distinguish between special skills and special accomplishments. Those with special skills perform at a high level at challenging tasks and win plaudits for doing so. Special accomplishments, by contrast, are accomplishments that are meant to look impressive and confer bragging rights. “I did an internship at Senator So-and-So’s office last summer.” “I worked in Nobel laureate So-and-So’s lab.” “I co-authored a peer-reviewed paper that was published when I was a high-school junior.” Such accomplishments often look impressive, but they can become far less impressive when the hidden costs to bring about these accomplishments are factored in (such as investments of large sums by parents to college consultants like Steven Ma). 

When it comes to special accomplishments, I would want these to make a difference in admissions only if it’s clear that they truly are a big deal — that they are hard won, gained without pampering or coddling, and satisfy objective criteria for excellence. Included here would be winning a big science prize, scoring high on an international math competition, being part of a nationally recognized dance team, belonging to a debate team that won a state or national championship, etc. 

I don’t want to exclude here students who in high school publish peer-reviewed papers or start non-profits to benefit groups in need. But with such special accomplishments, it’s crucial to make sure that the admissions committee is not being gamed. Such students need to make crystal clear that their accomplishments stem from deep conviction and are hard won. Conversely, it needs to be crystal clear that their accomplishments are not merely expedients for getting into an elite school.

The Prospects of Success?

Will this approach of using entrance exams as the main determiner of admissions to elite colleges and universities succeed? Obviously, to succeed it needs to be tried. By success, I mean making admissions to elite schools significantly fairer than they are now. Under success I would also include side benefits, such as curtailing the crazy arms race where students increasingly put extreme pressures on themselves to get into these elite schools.

I would expect the use of bespoke entrance exams in admissions to elite schools to lead to many more low- and middle-income students being admitted to them, especially if these schools continue, as they do now, to provide generous financial assistance to students and families from lower income demographics. Additionally, as just noted, this use of entrance exams would deincentivize the striving and self-immolation that now seem so much a part of getting into elite schools (the arms race). 

That said, I know of no elite school that on reading this proposal would smack themselves on the forehead and say “Of course, this is the approach we should be using in college admissions.” Perhaps the biggest obstacle for getting elite schools to make entrance exams the centerpiece of their admissions process is that schools revel in their low admissions. Low admissions confers bragging rights. “Harvard admits just over three percent of applicants.” “That’s nothing: Caltech admits just over two percent.” 

It’s a badge of honor among elite schools to be so desirable that applicants flock to them — the more the merrier. These schools can then, with seeming heavy hearts, inform the vast majority of applicants that, regretfully, they could only admit a very limited number of applicants because there were simply too many that applied. This is a sham. It perpetuates the unfairness of admissions to elite schools by excluding applicants who can do the work, but didn’t have all the breaks to make themselves look good on paper.

Even though elite schools have till now had little incentive to implement the entrance exam approach to admissions that I’m advocating in this essay, given the wild ride that higher education seems to be on now with the current administration in Washington, this entrance exam approach may not be entirely off the table. I could see, for instance, the current administration in the White House making federal grants to elite (and other) schools contingent on fair admissions, which could be defined as any students clearly capable of thriving at the school needing to have a fair shot at gaining admission. 

Taking this definition of fairness to its logical conclusion could mean randomizing admissions — essentially making it a lottery for students to gain admission once they demonstrate a certain threshold of ability and merit. I could even see an executive order enforcing such a lottery. In general, people rebel against lotteries where they want to win something that they think should not be subject to chance — such as gaining admission to an elite school. An entrance exam approach to gaining admission might then seem less arbitrary than a lottery while at the same time keeping applicants enthusiastically engaged with the school for which they are making the time to take a bespoke entrance exam. 

In sum, bespoke entrance exams for gaining admission to elite schools promise three benefits:

• They would improve fairness of admissions by leveling the playing field between privileged and less privileged students while also ensuring that only competent students are admitted.
• They would reduce, and ideally eliminate, the arms race among students vying to get into such elite schools.
• And finally, they would increase the likelihood that students gaining admission are there because they really want to attend the school and are willing to make the time to take such entrance exams.