I was just wondering if our undergraduates at Indiana University's Kelley School of Business were smarter on average than Harvard undergrads. They’re not, as I’ll explain later, but the comparison is not a ridiculous one—we’re really quite close. Full disclosure: I'm a Yale man with a doctorate from MIT (though I've also spent a couple of years as a visiting scholar at Harvard and like its current administration much better than Yale's), so Crimson misfortunes do give me a certain small pleasure.
When I was on sabbatical at Harvard, I volunteered to teach my regulation course for free, since I was curious what teaching Harvard students was like. (I'd been a TA as a grad student at MIT, but MIT undergrads live in a different dimension.) I didn't have to teach the course very differently. I thought the Harvard students were like the top third of the Indiana students, more curious and more willing to come to office hours, but with worse knowledge of business topics like present discounted value. But those were mostly economics majors at Harvard, who are probably smarter than their average undergrad. So I decided to figure out the numbers, which is not as easy as you might think, since Harvard doesn’t disclose its average SAT score.
Harvard says its 25th and 75th percentiles are "about" 1470 (98) and 1570 (99). This score is the sum of the math and verbal tests. I’ll add the national percentiles in parentheses so you can interpret the scores better. Thus, what these numbers mean is that 25% of Harvard students scores lower than 1470, but 98% of test-takers nationally scored lower than 1470. But this does not tell us the average score at Harvard. The average score is not the average of the 25th and 75th percentiles, because the distribution is skewed to the left, to low scores. You can't get higher than 1600 (99), the perfect score. On the other hand, below the 25th percentile the possible scores go down all the way to 400, and disclosed data does show that Harvard admits students with scores as low as 660, probably because of athletes, affirmative action, and alumni. Scores in the 600's are rare, but it looks like there are a lot between 1140 and 1380. Indeed, the obvious reason why Harvard doesn’t publish its average, just its 25th and 75th percentiles, is because the scores of the special-preference students below the 25th percentile, what we might call its “nonacademic admits” are embarassingly low.
Just this week, new information about Harvard is coming out in its trial for discriminating against Asian applicants. There are two lists of applicants who get special attention, the "director's list" and the "deans' list". These are for special friends of the university, which includes big donors. The Dean of the Kennedy School of Government wrote, “Once again you have done wonders. I am simply thrilled about the folks you were able to admit...[Redacted] and [redacted] are all big wins. [Redacted] has already committed to a building.” This is a sizeable proportion of Harvard students: "Documents published in June revealed that 192 current Harvard seniors — more than 10 percent of the Class of 2019 — are members of the dean’s list or the “Director’s List,” a similarly select group of applicants compiled by Director of Admissions Marlyn E. McGrath ’70 every year." It's a fair guess that quite a number of those who get in with SAT's below the 25th percentile are from those lists.
Incidentally, the Crimson reports: "Students who appear on the two lists — along with legacy students, recruited athletes, and children of Harvard faculty members — are overwhelmingly white. Students in these four categories make up roughly 29 percent of students at the College." Whites are 44% of Harvard students. Thus, it seems that if one excludes the special list students, affirmative action students, and Asians, there are only 15% of slots left over for white students who are not alumni kids or athletes. The great bulk of white students who apply have little chance of getting in. From a different angle, Norwegian-Americans and German-Americans (to name two groups that interest me personally as much as half-Koreans do) are probably rare birds at Harvard. It would be interesting to know what fraction of Harvard students are white but of neither Jewish nor Puritan descent.
Back to business. Let's try to estimate the Harvard average SAT score. 25% of its students are between 1570 (99) and 1600(99), so let's assume the average of that group is 1585(99). 50% are between 1470 (98) and 1570 (99) and let's assume the average of that group is 1520 (99). 25% are between 660 (1) and 1470 (98), and let's assume--- the big assumption-- that the average of that group is 1340 (90). The grand average is then .25(1585) + .50(1520) + .25 (1340) = 1491 (98).
The average score at Indiana’s Kelley School of Business is 1390 (94). For direct admits (as opposed to those who are admitted in the sophomore year after being at IU and getting good grades their first year), it takes a minimum score of 1380 (93) plus other qualifications, and their average has been rising sharply and was 1437 (96) for Fall 2018. Note that there was some kind of re-norming going on, too. If someone sees that this interferes with my analysis, please let me know, and I'll revise it.
Thus, in the end I think Harvard students are indeed smarter on average than my Indiana students. For Kelley students to be better than Harvard ones, the bottom 25% at Harvard would have to have an average of 1130, which is unlikely. Still, Kelley is pretty close, and our students have been steadily improving over the past twenty years, with no sign of a slowdown. Indeed, the improvement has accelerated over the past five years. We can hope to catch up.
How about some other schools? Indiana University-Bloomington overall, not just business majors, has an average SAT of 1276 (84), which, like Kelley’s, has been sharply rising in recent years. My daughter is at Hillsdale, where the average SAT is 1350 (91), lower than Kelley. Her SAT was something like 1500, but she did not plan on studying business, and we decided that Hillsdale is tops for liberal arts, with Baylor's Honors program second. Pennsylvania’s Wharton School, the top undergraduate school for business, has an average SAT of 1499, just a little higher than my estimate of 1491 for Harvard. Washington University-St. Louis, at 1507, has the highest SAT of all undergraduate business schools.
Dartmouth College is an Ivy League whose students seem to be close to Kelley direct admits. 25% of Dartmouth students score between 1560 and 1600, and let's assume the average of that group is 1580. 50% score between 1350 and 1560 and let's assume the average of that group is 1455. 25% score between 930 and 1470, and let's assume--- the big assumption-- that the average of that group is 1340, the same as Harvard’s lower tail. The grand average for Dartmouth is then .25(1580) + .50(1455) + .25 (1340) = 1458, which compares with 1437 for Kelley direct admits.
Purdue surprised me. Like Harvard, the university as a whole only reports the 25th and 75th percentiles, which are 1160 (68) and 1360 (92). Testive.com says Purdue's average is 1195 (74), and Prepscholar.com says it is 1300 (87), but websites often misreport SAT scores--- because, I think, of muddle over what “average” means. Purdue's College of Liberal Arts has an average of 1142 (65). I bet they want to advertise how low it is so as to attract majors. Purdue Engineering doesn’t disclose its average, but its 25th and 75th percentiles are 1320 and 1440. Recall that Kelley’s average is 1390 (94) and its direct admits average is 1437, so it looks like Indiana Business students are considerably smarter than Purdue Engineering students.
References I didn't get round to putting as links in the main text: