An admissions lottery?

Chad Aldeman makes the case in Inside Higher Ed today that selective universities should replace their current admissions systems with a simple lottery. The idea is to establish a threshold level, based on standard criteria like high school GPA and SAT scores, above which all students will have an equal chance at admission through a randomized drawing:

At many institutions, in other words, it is a far more random process than colleges would like students to believe. The myth of a meritocracy, on which the selective admissions system is built, is substantially a lie.

Selective colleges did not mean for this to happen; rather, they are victims of their own success, along with the emergence of a truly national higher education market and the rise of a rankings-driven consumer culture. But, there is no going back now, so colleges should embrace the unavoidable randomness and go from a lottery-like system to a true lottery.

Institutions would set a threshold based on high school grades and SAT score and then open the lottery to anyone meeting those levels...

This idea is so bad on so many levels that it's hard to know where to start. First let's consider how the threshold will be set: Aldeman suggests that it should be a function of grades and SAT scores. Of course, making it a function of "grades" won't work, because this will virtually eliminate the incentive for high schools to give any grade less than A, and transcripts will become little more than worthless pieces of paper. (This will be especially true at the richest and most college-oriented schools, whose students are presumably not the ones Aldeman hopes to benefit.) You can correct for this by making the threshold a function of class rank instead, but this merely leads us to a different set of adverse consequences: students will have an incentive to attend the shoddiest and least competitive schools they possibly can. This would be an interesting adventure in egalitarianism, but I don't think it should become the basis of education policy.

Second, Aldeman oversells the extent to which admissions is a "random" process. No one can deny that there is a tremendous amount of luck involved in the process for applicants at the margin, whose credentials are neither so strong that they merit clear acceptance nor so weak that they demand automatic rejection. Admissions officers do not have the time necessary to make a full, reasonable evaluation of applicants' potential. This is, however, still far from a completely random process. The existence of a sizable random component in admissions decisions doesn't mean that there's no legitimate component, or that coin flipping would be just as effective.

Consider one piece of evidence: the results from last year's Putnam math contest. Out of the 74 top finishers listed on the results sheet, I count only 15 that are not from MIT, Harvard, Princeton, Stanford, Waterloo, Toronto, Caltech, or Duke. That's almost 80 percent concentration in eight universities—28 percent at MIT alone! Clearly admissions offices are doing a little more than "random" selection here. You might say, reasonably, that this is a special case, because success on the Putnam math contest is highly correlated with success and experience in high school math contests, which provide an easy and accessible measurement tool for admissions officers. But then we're admitting that there is a useful, objective tool beyond GPA and the SAT that guides colleges to productive admissions decisions. Should we include an adjustment to our threshold for success in math contests? What about the myriad other qualifications that are also useful predictors of performance? If we do, we gut the simplicity and transparency that Aldeman finds appealing in this system. If we don't, we're relying on a crude SAT cutoff and random number generator to match our most talented students to the appropriate universities.

Finally, you have to wonder why any school would commit itself to such a policy. If a lottery threshold of X will make Harvard receive 100 times as many applications as spots, what's to stop it from raising the threshold to X+1? This would raise the average quality of its students with absolutely no cost, except possibly forgone lottery fees. Indeed, why wouldn't Harvard want to raise the threshold so high that there's no lottery at all? You need to have a tremendous amount of faith in universities' commitment to the social goals of the lottery system (whatever they might be) to assume that they will refrain from taking a costless, painless measure that is guaranteed to produce a more qualified group of students. Even if there are societal benefits to Aldeman's system -- and I don't think there are -- the return to individual universities from defecting are simply too high for it to be stable. All it takes is one good institution to decide against a lottery, or a gradual upward creep (motivated by institutional self-interest) in thresholds until the "lottery" no longer exists.

All in all, this is a terrible policy proposal, one that will not be implemented and would produce disastrous consequences if it ever was. It's curious, in fact, how a seemingly smart and perceptive person like Chad Aldeman managed to conclude that this was a good idea. I blame it on the perverse appeal of counterintuitive thinking. The notion that we should raffle off college admissions seats sounds crazy, and if we convince ourselves that it's actually well-founded, we'll pat ourselves on the back for our brilliant analysis and ability to transcend mere intuition. The only problem is that it actually is crazy.