Send your kid to Harvard
contra Nate Silver
Nate Silver has advocated for going to a state school over an Ivy for several reasons:
People’s perceptions of higher ed are declining.
Public opinion on elite universities is worsening, particularly for Republicans.
Ivy students potentially benefit from grade inflation, coddling, and dubious admission policies.
He is correct in saying that American perceptions of higher ed are declining, but is mistaken in thinking that the value of a university degree is built on stated public preferences, it’s by what students signal by graduating. While the average IQ of a college graduate has been declining, educational attainment still signals intelligence well, so rational employers will still value it.
Nate Silver is probably correct about Ivy league students are more left wing than normal college students — but if they’re more intelligent and competent, why should this matter at all?
Compare this to Harvard, where the average SAT scores of the applicants have stayed consistently in the top 1% for 50 years: 1508 (99.8th percentile) in 1970 and now 1520 (99.5th percentile) in 2024. Assuming a .8 correlation between SAT scores and IQ, the average Harvard student should be expected to have an IQ of 135 in 1970 and 131 in 2022. Empirically, Harvard students in 2003 scored an average of 122 on the WAIS-R after adjusting for the Flynn Effect, though the sample was small (n = 86) and 61% Female.
Top universities in the USA have similar average SAT scores (the lowest Ivy being Cornell at about 1470), so the average IQ of an Ivy leaguer is probably only a little lower (say, 128).
While the relationship between college selectivity and earnings is confounded by intelligence, this is not the case at the tail end of college selectivity. Regardless of whether admission controls or value added estimates are used, Ivy+ (Ivies, MIT, Stanford, Duke, UChicago) schools have a causal effect on income, based on waitlist decisions. While the waitlist decisions are not random, they are made on factors that are not related to income, so they can be used as a causal estimate of the effect of admission on life outcomes.
The logic of focusing on waitlisted applicants is similar to that underlying a regression discontinuity design: waitlisted students are close to the margin of admission and may have similar potential outcomes, potentially permitting identification of causal effects of admission by comparing the outcomes of those who are admitted with those who are not. However, since waitlisted applicants are not admitted randomly, there is no guarantee that those who are admitted from the waitlist have the same distribution of unobservables correlated with outcomes as those who are not.
We therefore begin by evaluating whether the variation in admissions decisions among those on the waitlist is driven by idiosyncratic factors that do not affect outcomes or systematic factors that do using the multiple-rater admissions test developed above. Formally, we treat an indicator for being placed on the waitlist as an observable control and test whether the residual variation in admissions conditional on being on the waitlist at a given Ivy-Plus college A is correlated with admissions outcomes at other Ivy-Plus colleges B.
…
In sum, the balance tests show that admissions from the waitlist are non-random, but are driven by idiosyncratic factors orthogonal to potential outcomes, consistent with the results of our multiple-rater tests. We therefore proceed to compare the outcomes of students accepted vs. rejected from the waitlist using the estimator in (5) to identify the causal effects of admission
I also object to the notion that public approval of these institutions is highly important. Consider Unz’s 2012 piece The Myth of American Meritocracy, which found that Jews are vastly overrepresented at Ivy league universities in comparison to what would be expected from their average ability. Did this dent the reputation of the schools? Well, we are still discussing the potential “coming fall of the Ivies”, so I’m inclined to think that it didn’t.
I assume that it would still be possible for a shift in public opinion to damage these universities — imagine if the true average IQ of a modern Harvard student was 113 and somebody tested their IQs using the WAIS and posted the results. That would be a pretty big hit, and if subsequent testing found the same thing. Or that all the Ivy league presidents met up in a secret building and discussed how to keep Jews and Asians out of the schools. But at the moment, the deadest skeletons in the Ivy league closet are admissions methods that gate out Asians (which I support) and grade inflation (school is signalling, so I support it).
II.
Let me try to make a more intuitive case for why going to an Ivy league school matters.
It’s difficult to make causal inferences about this data, as Ivy-Plus graduates are already selected for cognitive ability (mean IQ of ~128) and presumably other traits as well. If the entire 18 year old population (4.4 million) of the United States who have IQs of below 121 were killed, this would leave a population of 352,000 (4.4M * .08) with an average IQ of 128, roughly matching the Ivy+ average. Compare this to the 21,961 students in the Ivy-Plus class of 2027 — the probability a highly intelligent person attends an Ivy+ uni as an undergrad is only 5.6% — lower if you consider in. Assuming some non-negligible overlap between undergrads and postgrads, the percent of highly intelligent Americans who graduate from these institutions should only be about 7-9%.
To explain the discrepancy between the expected and observed rates of representation of Ivy+ leaguers in the elite, two alternative hypotheses must be considered:
Ivy+ leaguers are favoured over normal college graduates independent of other traits.
Ivy+ leaguers surpass normal college graduates in traits beyond intelligence.
Regardless of which of these hypothesis are true, or more true: it’s better to go to an Ivy league institution than a normal university. If Ivy leaguers are somehow selected for traits beyond intelligence, then that will make them even more valuable to employers. If the first theory is true, then it would be ideal to go to an Ivy to benefit from the signalling value.
If graduates of highly selective colleges are given an abnormal premium in wages and status, then it would be economically rational to select against these degrees in the presence of an abundance of information, but this strategy is still compatible with prioritizing attending a highly selective university on a personal level.
This is probably more appropriate for general liberal arts. By the time you dive into some of the STEM fields the advantage of the ivy education strikes me as less certain - at least with respect to flagship state universities, which frequently have departments with higher reputations than the corresponding departments in most of the Ivy's. For example, the University of Washington's CS program is more highly rated than that of most Ivy's. I have no doubt that the cachet of Harvard or Yale will make a difference for selection of a small number of people into executive programs but you have a lot of people making enormous investments for a very low probability payoff. The vast majority of the people in STEM will do about as well with an education from a solid state university as they would have with an education from an Ivy.
I did my Physics BS from the University of Maryland, College Park more than 50 years ago. The college was open admission. So was the Physics department. We started with something like 200 students in the Physics for Physcists classes. At the end of the Sophomore year there were 5 Physics majors left. The next year we picked up a few transfers. It is my understanding that Math had a similar filter function. The fact that it was easy to get into the University said nothing about the demands and expectations of specific departments - which could and did enforce their own standards, so an analysis at the University level may not provide significant insight into the characteristics within fields within the university.
Did you use a pic of Yarvin because he looks dashing in those shades or because he went to Brown?