For and Against Lotteries in Elite University Admissions

People (whose names often rhyme with Palcolm Sadwell) sometimes suggest replacing the admissions systems of highly selective universities with lotteries. The proposal is that universities would mark a pool of students as ‘good enough’ and then students from that pool would be accepted at random. Here are some arguments for and against this idea, inspired by Julia Galef’s unpopular ideas series. I don’t agree with all these arguments (or even fully endorse any); the point here is just to categorise the arguments worth considering on both sides.

For:

  • Competitive university admissions generate enormous amounts of waste (through Goodhart’s law) with time and money being spent on extracurriculars and private tutoring for what is essentially a zero-sum game. It may also be incentivising unethical behaviour: the proportion of students who admitted to cheating in high school went from 34% in 1969 to 74% in 2002.

  • Current admissions are racist and classist, especially when they rely on interviews, because admissions officers are unconsciously biased.

  • Noise is systematically much higher than people suspect so the system is already, to a moderate extent, random.

  • With a lottery, admitted students would be from more racially and socioeconomically diverse backgrounds. Different countries and universities have differing levels of remedial discrimination policies, so a lottery would make some colleges less diverse, but the net effect would be to make them more diverse.

  • The current system penalises students who spend their teen years doing things they actually enjoy. There’s adverse selection for people who are unoriginal, uncreative, and place little value on their time.

  • Students applying to top universities are very stressed out, e.g. 46% of Cambridge students either have depression or suspect they might. Easing the competition would help with mental health problems and lead them to put less pressure on themselves.

  • If getting into top universities is more arbitrary, it will cease to be a reliable signal of competence to employers. Therefore, the labour market will rely more on non-credential signals of competence, bursting the higher education bubble.

  • Lotteries will make universities more intellectually diverse. Conservatives are highly underrepresented in top universities, and (at least in America) this is in part because of ideological admissions officers that discriminate against people they disagree with. More generally people are biased in favour of their in-group. Lotteries give a boost to the academic careers of heterodox thinkers, who can push back against echo chambers that produce low-quality research.

  • Grades and chances at admission do not signal how much people will contribute to society. For example, a study at Michigan Law found that black students admitted through affirmative action had worse grades but contributed equally (measured by satisfaction, income, and ‘service contributions’) in their later careers.

Against:

  • Thinking that students get into top universities because of privilege confuses correlation and causation. Private schools do not particularly benefit children academically, and standardised tests are very hard to game. Thus, what looks like bias in who gets admitted is actually a statistical illusion.

  • If the problem is Goodhart’s law leading to gaming of the wrong metrics, you can solve this with an admissions system that is opaque from the outside, or with tests that are difficult to predict in advance (Oxbridge interviews are famously difficult to prepare for).

  • Many university admissions systems (like Ireland’s) already work well and don’t waste students’ time without recourse to a lottery. Even Oxford and Cambridge accept a relatively high proportion of acceptable candidates. The problem of zero-sum competition over irrelevant metrics for university spots is exclusively faced by a small number of (mostly privileged) Americans. The people who say this is a niche problem are correct. It appears to be a much larger problem than it is because the people who went through this gruelling process are disproportionately represented in academia and the media.

  • A few elite intellectuals make almost all the contributions to academia, attending a prestigious university is important to their success, and yet they would have the same chance as anyone else ‘good enough’ under a lottery system. Since these people are so socially valuable, it may even be worth some social injustice in higher education, in general, to make sure we catch every genius with a high chance to make a large contribution.

  • A lottery isn’t fair. Someone who worked extremely hard on their application would have the same chance as someone who applied as an afterthought and was barely good enough.

  • A lottery undercuts the mythology of education. We live in a society that values individual achievement and treats people as solely (or mostly) responsible for their own success. This may not be the best way to organise society, but it is currently the case. Changing to a lottery would lead to the winners feeling undeserving, and the losers feeling cheated.

  • Admissions are not actually very biased. The research on unconscious bias is plagued with problems, and most of the studies referenced in books like Blink have failed to replicate. People in charge of admissions are competent professionals who make their decisions on basically sensible grounds.

  • Less privileged and wealthy students mean that the university will get fewer donations. While it is perhaps sad that universities need to rely on alumni donations, they also do a large amount of social good, funding important research and scholarships.

  • Applying to university is the first stressful interaction many secondary school students have with the real world, and it teaches them many valuable skills: how to prepare for interviews, how to write good cover letters, and how to deal with rejection. This is only effective if students are trying their hardest and the admissions procedure isn’t random.

Thanks to Gytis Daujotas, Cian Mullarkey and Luise Wöhlke for reading drafts of this post.