On Social Verifiability
Posted: 2023-05-07 17:57, last updated: 2023-05-15 17:44.
Like law, standards have an expressive function. What then are we to conclude from the fact that admissions standards at many prestigious New World institutions are no longer communicated? We must conclude that the new standards are literally unspeakable.
The abolition of transparent admissions standards, such as SAT scores, has not led to their replacement by another clearly defined numerical scheme. Rather, individual bureaucrats have been empowered to make up standards on the spot. Everyone else is left guessing. Opaqueness breeds suspicion, and a lack of what I call social verifiability.
Arguments in favor of non-transparent standards for admission to college are often based on a critique of meritocracy. And indeed, merit and meritocracy are ill-defined, empty concepts. Meritocracy begs the question of what standard should be used to ascertain merit. We do not know answers to these questions. Rather, meritocracy appears to be used as a rhetorical device to smuggle a certain viewpoint of merit into the debate. Although that will typically be my viewpoint (see below), “meritocracy” should be rejected because it is not rich enough as a concept. It does not stand on its own. What proponents of non-transparent standards forget is that meritocracy admits any standard, and that their program is indeed also a form of meritocracy, merely with more or less secretly redefined merit.
Standards are crucial to delineate because they change behavior. The best response to a clearly communicated objective standard is to optimize for some target. If standards are opaque, individuals have to optimize in a variety of dimensions, but without clear guidance. Admission becomes arbitrary, checks and balances are removed; reconsideration and review become impossible. Trust in institutions requires the absence of black boxes.
The fact that some objective standards are easily gamed should detract only from those easily gamed standards, not from objective standards writ large. For example, recent advances in the field of large language models make essays unsuitable for admissions; more importantly, the desired political bent in these essays is now not guaranteed to be genuine. SAT scores are emphatically not easily gamed. As standards, they have been purposely constructed to be purely knowledge- and skill-based. Some knowledge and skill are necessary to master the SATs. Which other standards are not easily gamed, too, is an interesting question for future research.
Societal ills such as discrimination may be present in the SAT score. However, the SAT by itself is completely (facially) neutral. Differences in outcomes must result from prior inequalities, such as in access to education. There is no issue addressing those at the admissions level, but the fact that adjustments are kept secret demonstrates that the adjustments are not obviously benign. Opaqueness is used to discourage and complicate public scrutiny over admissions. Intransparency cannot be to the benefit of the public.
Civil law systems are often criticised by Americans because a single government employee (the judge) typically has the power to convict a defendant and order punishment. By itself, such a system is obviously inferior to a jury system. But what highly developed civil law systems get in exchange is social verifiability. Juries' decisions are exactly subjective; appeals courts have to mindread whether any piece of evidence or instruction may have improperly influenced their judgment. In a civil law system, judges have to clearly spell out their reasoning; they cannot hide so easily. An appeals court can verify that reasoning, and so can anyone else. There might still be disagreement, but the innerworkings of the system are subject to external control. Its mechanisms can be examined on the level of an individual case, making the judgment informationally richer. The judge is forced to apply a minimum degree of logic and common sense; no such things are expected from jurors. Because juries' decisions cannot be rationally examined, they are processually inferior to a bench trial with a judge who has undergone years of training. It is no wonder that is has been claimed that an innocent defendant would prefer a bench trial, whereas a guilty defendant would prefer a jury trial.
And it is for that reason, too, that civil law systems fare poorly with respect to questions of the admissibility of evidence: The effect of a piece of evidence on a judge can be theoretically traced, whereas some evidence must be—paternalistically—withheld from the jury because its members are not sufficiently sophisticated. The “answer” of American political economy to strong protections for defendants on the extensive margin—often wholly invented from wholecloth by the Supreme Court—has been to blow up the intensive margin. This is not obviously an improvement.
The strict formalism in the civil law system can also open it to abuse, such as for petty crimes. I have argued elsewhere that one great issue in civil law systems is the lack of jury nullification (although nobody can be finally convicted in Germany without both people's representatives agreeing, a little-known fact), and that civil law judges must be given the unappealable power to dismiss charges. Lopsidedness can help solve the injustice of rigidity; but only in criminal law, where false positives must be avoided. In admissions, capacity constraints make some ranking unavoidable.
An absence of objective standards removes control and accountability. It also makes planning much more difficult and adds needless layers of surprise. It is needless to say that admission (or not) does not make a human being more worthy than another. But there is a value judgment inherent in admission (and the sentencing of a defendant). Its value derives from being rational, from being objectively verifiable and derived from facts and logic. Anything else sounds just like someone's opinion, man. Standards cannot be known to be just if they cannot be known. The paternalistic idea that bureaucrats just know better must be rejected if they refuse to show their work.
Even if standards are made to be fully transparent, I argue that they should be also be as simple as possible.
The German system is practically optimal in this regard. Undergraduate admissions generally work as follows: Every applicant's high school grades (Note der allgemeinen Hochschulreife) are digitally recorded (actually just a single number per applicant). The university decides based on practical and statistical considerations how many places to offer. Applicants are then ranked according to their magical number, and there is a cutoff (the "numerus clausus") somewhere if the course of study is popular enough. The end. There are no shenanigans. Graduate admissions follow the same general pattern, but there might be other factors as well; typically, all or almost all factors are objectively verifiable. In any case, no consideration is ever given to immutable characteristics.
This simple standard allows students to focus single-mindedly on high school studies. Students know what is expected of them. It is simple to see that focusing on high school performance is also just. Performance in other dimensions is highly correlated with high school performance. It is a common error to assume the existence of multiple intelligences; it's pretty much all the same thing. Schools are supposed to give students the most basic knowledge required to function in society. If we detract from this important desideratum by rewarding extracurricular activities, we cannot be surprised when students cannot read. There are no incentives to focus on the essentials. I believe it is deeply injust to burden students with extracurricular activities that are technically irrelevant—because they correlate highly with grades—and practically detract from actual scholastic achievement. Americans ask both too much and too little of their students.
Undergraduate studies are, controversially, a continuation of high school. For that reason, high school grades should be determinant for admissions.
Simple standards such as grades also allow disadvantaged individuals to theoretically transcend the path preset for them. I can speak from my own experience that playing the violin or lacrosse would not have been possible for me financially. These items would not have identified me, and I would have had no interest in them. Grades may be practically unjust, but extracurricular activities are both theoretically and practically unjust. Every human being is fundamentally equal in the sense that we all have 24 hours per day. What happens in our mind is the only thing that definitely belongs to us and cannot be taken away. Modern technology grants us the opportunity to further our horizons and to learn, even to those most impoverished. This is not true for things outside of our mind. Alternative standards are not necessarily more equal and may be much worse, especially when combined.
Fundamentally, simple standards are also easier to validate. In a metaphysical sense, they are more “transparent” even if they are as open as a complex admissions algorithm. A simple standard may be less just, but its simplicity lets no one forget that admission really is not about human worth or the lack thereof; it's just business, it's nothing personal. Simple standards grant the affected individual a form of reprieve in which they are freed from the suspicion and discomfort of rejection: What could I have done differently? Simple, transparent standards are a rejection of societal “justice” that is only ever evaluated statistically, and an embrace of personal, individual justice, respect and equality before the provost. They eviscerate mind games, or the suspicion thereof on the part of others.
Who is to be admitted to a prestigious university is not an easy matter. There is no universal understanding of what is fair or just. We will not micromanage our way to justice. Clear and simple objective standards for admission to university communicate what is desirable. What is desirable should be discussed and deliberated upon publicly; it should not be left to unaccountable bureaucrats. The fact that standards are increasingly opaque shows that they must be repugnant. While merit and meritocracy are both empty constructs and rife for exploitation, simple standards could help students refocus on school achievement, removing unnecessary and wasteful distractions. Simple standards such as SAT scores are profoundly just because they give clear targets and paradoxically remove the impression that a rejection from a college may be a personal slight.