Let me try again to explain what Stephen Pinker gets wrong about admissions at places like Harvard and Williams. Outside of academics (grades/scores), sports (that the college plays and recruits for), race (mostly black but also Hispanic) and wealth (mostly being poor but also some super-wealthy), virtually nothing else matters to elite college admissions. This is the central thing that is wrong with Pinker’s description — with its talk of “the arts, charity, activism, travel” — and even with the understanding of sophisticated observers like Eric Rasmusen and Stephen Hsu. Consider Hsu:

Some have quibbled with Pinker’s assertion that only 5 or 10% of the Harvard class is chosen with academic merit as the sole criterion. They note the overall high scores of Harvard students as evidence against this claim. But a simple calculation makes it obvious that the top 2000 or so high school seniors (including international students, who would eagerly attend Harvard if given the opportunity), ranked by brainpower alone, would be much stronger intellectually than the typical student admitted to Harvard today.

First, it is somewhat rude of Hsu to use the phrasing “some” when I (I think) have been the only one to make this argument explicitly.

Second, by “quibbled,” I think Hsu means “demonstrated.” Where, exactly, is the flaw in the argument? A 1590 Math/Reading SAT score is far into the 99th percentile (pdf). (Exactly how far is unclear to me.) Here (pdf) is data from 2008 that uses all three sections and here (pdf) from 2013. The total number of perfect scores is in the hundreds not thousands. I realize that I am playing fast and loose here, that I began by looking at Math + Reading, that these tables use all three sections and that summing percentiles is not the same as looking at the percentiles of the distribution of summed scores. But the basic numbers make clear that there are not nearly as many high scoring students at Hsu naively assumes.

Imagine that the height of the 75th percentile male student at Harvard were in the 99th percentile of the overall distribution. That would be about 6’4”. If 1/4 of the men at Harvard were above 6’4” would you doubt that height was the central factor in admissions? There is no way to get so many tall men if you are selecting on other factors that are not heavily correlated with height.

Third, Hsu admits in the comments that his “simple calculation” is off by an order of magnitude! This is the nice thing about being a theoretical, rather than experimental, physicist! Getting the answer within one or two orders of magnitude is close enough! ;-)

There are not 2,000 high schools students that are much stronger than the typical Harvard student, there are 200. And someone has to go to Yale!

Fourth, Hsu uses phrases like “ranked by brainpower alone” (by which he means a heavily g-loaded IQ test) and implies that this is the same thing that we (me and Pinker) are talking about when we discuss “academic merit.” But it isn’t. “Academic merit” means the highest grades in the most rigorous high school classes along with top teacher recommendations and extreme standardized test scores. This is what Harvard (and Williams) care about. You can be a genius but, if you blow off high school classes that you find boring and stupid, Harvard/Williams don’t want you.

The central issue here is not: What should Harvard/Williams select for? (I expect that Hsu and I would see eye-to-eye on that.) The issue is: What do Harvard/Williams select for? I am trying to explain that Harvard/Williams do not care nearly as much about “extracurricular” activities — at least things like “the arts, charity, activism, travel” — as Pinker/Rasmusen/Hsu think they do. The data demonstrate this because there is no way for Harvard to have such extreme SAT scores and, simultaneously, place much/any emphasis on these other factors. If it did, then, almost by definition, the SAT score distribution would be shifted lower.

To be concrete: Harvard admitted 2,000 high school students out of the 35,000 who applied last year. Imagine that we kept fixed the things that I say matter (academic merit, race, wealth and recruited athlete status) and randomize the things that Pinker/Hsu think matter (arts, charity, activism, travel). That is, Applicant X keeps the attributes that I assert Harvard cares about but then is randomly assigned the extracurricular activities of some other applicant. How different would the admissions decisions be?

Almost indistinguishable! Harvard would let is at least 1,500, and probably more like 1,900 of the 2,000 students it did, in fact, admit even if they were assigned a random set of extracurricular activities from among the pool of applicants, and for all the reasons that I gave in my previous post. These facts turn Pinker’s claim on its head. It is not 5% of the class that is selected on the basis of academic merit. It is 5% that is selected on something other than academic merit, once we control for race/wealth/athletics.

Again, if Pinker had just said that race/wealth/athletics play a big role, and stopped there, I would have no complaint. They do! But Pinker misleads his readers, and people like Hsu/Rasmussen with his talk of “arts, charity, activism, travel.” That stuff plays no material role. In fact, one reason Harvard likes to act like those things matter is so that it can hide the big influence of race/wealth/athletics behind a patina of “We care about great violin players too!”

How to convince Pinker/Hsu/Rasmusen of this fact? Consider Caltech. Hsu suggests (and I agree) that Caltech bases its admissions standards almost completely on “academic merit.” Consider the 75th percentile of its score distribution (pdf):

Now Harvard (pdf):

At 75th percentile of the SAT/ACT, Caltech (800/800/790 and 35/36/35) is indistinguishable from Harvard (790/800/790 and 35/35/35). If we assume that Caltech admits on the basis of “academic merit” (it does!), then it must be the case that Harvard uses, more or less, the same criteria (at least for 25% of its class), otherwise, it would not have the same extreme distribution of scores. (And even the tiny advantage to Caltech is probably explained by Harvard putting more emphasis on the high grades portion of academic merit than Caltech does, not on the (imaginary!) extra bump that Harvard gives to excellent sculptors who score 2390 over average sculptors who score 2400.)

Now, obviously, below the 75th percentile, things change. But that is not because Harvard cares about the arts, it is because Harvard cares about race/wealth/athletics. Key statistics:

Black students: Harvard (7.0%) and Caltech (1.6%)
Pell Grant Eligible students: Harvard (17%) and Caltech (11%)
Division 1a caliber athletes: Harvard (12%) and Caltech (<1% ?) I lack the energy/data to go through the argument that these factors alone explain the differences in the 25th percentile but they almost certainly do. But, before going further, can we at least agree that I have demonstrated that, at least for 25% of the class, Harvard is at least as focused on “academic merit” as Caltech is? In other words, Pinker’s estimate of 5% or 10% is off by at least a factor of 2.5 to 5. Once we agree on that, we can move on to other portions of the score distribution.