Tuesday, August 15, 2006

MIT vs Caltech: Nobel count

[ 2015: Updated information. ]

I learned from the Caltech News alumni magazine that 17 Caltech alumni have won the Nobel prize, versus 25 from MIT. You might think the advantage here goes to MIT, but their student body is 5 times larger! Most people are shocked to learn that Caltech's graduating class is only about 200 students. On a per-capita basis, I believe Caltech produces more science Nobel prizes than any other school. Keep in mind that about half of Caltech undergrads major in engineering or computer science, which are not Nobel-eligible disciplines.

When I was a student we used to joke that MIT stood for "Many Incompetent Technologists" (emphasis on Many) or "Made In Taiwan" :-)

On the other hand, Feynman went there, so it can't be all bad. Actually I have to admit it is probably more fun to be an undergrad at MIT than at Caltech (part of it is that the classes are so much easier :-). When I lived in Cambridge I could see there was a much more lively college scene in Boston than LA. MIT is bigger and has a better male-female ratio and more balanced social life than Caltech. Of course, the climate in Boston isn't as nice.

Before I had twins and startups I used to be involved in Caltech admissions and recruiting -- including calling up admitted students to answer questions and give advice. I was often speaking to students who had been admitted to both Caltech and MIT, and I was always scrupulously fair in describing the pros and cons of the two places.

More Caltech bragging: patents and PhDs.


Anonymous said...

You might enjoy the intro to a talk I gave at MIT while I was a postdoc at Caltech. Best viewed with the animations : http://dabacon.org/home/talks/mit01.ppt

Anonymous said...

I once did a spreadsheet, Nobel laureats and the instutions where they got their undergraduate degrees only (not faculty or grad students), and compared the totals to the estimated number of undergraduates produced by each institution over the past 100 years or so. Caltech was way ahead of anyplace else for Nobels per undergraduate produced. University of Chicago was second place. Harvard, MIT, Princeton etc. were all clumped together within error bars.

Steve Hsu said...

Interesting data -- I would have guessed that the Caltech, Chicago and MIT curricula were most rigorous among top schools, and the students there among the most serious. I wonder if that plays a role in the results?

See my earlier post on revealed preferences and college rankings (search on the rhs of the page). Head to head (i.e., when a student is admitted to two places), Caltech does very well against other schools.

Anonymous said...

The Caltech undergraduate Nobel per capita is a great advertising tool and I've always been surprised that Caltech admisions doesn't use it more. Maybe it sounds too arrogant or maybe they think it might jinx Caltech (Of course I'm baised in all this, my grandpa went to Caltech!)

Anonymous said...

Nobel prizes are only offered in certain fields. There are no Nobel prizes for math or engineering. The MIT departments that match the fields where Nobel prizes are available are much smaller than the engineering departments. For example, one-third of the MIT undergraduates were EECS in the days when I went there. The only deparment in the school of science with more than 10% of the undergraduate population was Biology.

How does the comparison between student body size and Nobels change if the comparison is made using only departments in fields where the prizes are available?

Steve Hsu said...


The modification due to factors you mentioned is small. I would guess that EE, CS and E&AS (engineering and applied science) -- three majors that generally do not lead to Nobels -- account for almost 50% of all students at Caltech. (Last I checked EE and E&AS were the most popular majors, although physics was close when I was a student.)

I was discussing the data with some colleagues, and it is really striking. I would say that with high statistical confidence one could say that the Nobel output per unit time for each institution is similar -- almost certainly within a factor of 2. But, given the relative student population sizes that means Caltech outproduces MIT per student by at least a factor of 2 with high confidence. (The a priori probability that a Caltech student in a Nobel-eligible discipline will win a Nobel seems to be at least twice the corresponding probability for an MIT student, at high statistical confidence.)

An aggressive statistician like La Griffe du Lion (http://www.lagriffedulion.f2s.com/) would probably use the data to deduce how Nobel -level work is predicted by metrics like SATs and high school achievement. Caltech students tend to be about one sigma above MIT students on average. That means the high end tail (where, presumably, the Nobelists come from) might be populated very differently, thereby accounting for the difference despite the large overall population of MIT.

The calculation is analogous to one in sports, where some ethnic groups which are a small minority produce more elite athletes than the majority, even though the average differences are relatively small. But we can't talk about that here ;-)

Anonymous said...

Steve -

I wasn't aware that the engineering majors at Caltech were as popular as they are at MIT.

Looking back, MIT did offer different core undergraduate courses for students with different levels of ability. During the first week on campus, we all had to take a math test. The low scorers were assigned to physics courses that didn't require immediate fluency with calculus.

At the high end of the scale, I took physics sequence that assumed the students were already familiar with multi-variable calculus. There were only about 30 of us taking it. We had our labs run by Henry Kendall. A few years later, Wolfgang Ketterle lectured the course. It's not Feynman, but its not bad either.

Was there any tracking at Caltech?

Steve Hsu said...


There is tracking, but I think everyone knows calculus and the Ph 1 sequence assumes that. When I was there everyone (including biologists) had to take a pretty rigorous 2 year math and physics sequence, that included (among other things) multivariable calculus, ODEs and PDEs, linear algebra, statistics, Lagrangian mechanics, QM up to Schrodinger's equation and the H-atom (so, PDE based), stat mech including thermo, microcanonical and canonical ensembles, etc.

This is normal for most physics majors, but very exceptional for most others -- very few biologists at other schools would take that much math and physics, and even most math majors at other schools wouldn't take that much physics, etc.

Peter Shor (a Caltech undergrad), who came up with Shor's algorithm for factoring on a quantum computer, probably knew QM well because of the Caltech curriculum. (His PhD is, I believe, in combinatorics at MIT and he was at Bell Labs when he made his discovery.)

The curriculum is hard for a lot of people and the attrition rate is the highest among elite universities. The graduation rate was only about 80% when I was there.

Anonymous said...


As an undergrad at Chicago, I'm somewhat surprised that Chicago has produced so many outstanding undergraduates. My impression has always been that while U of C students are serious, (science) students at the Techs are more talented. I would be interested in seeing your spreadsheet, if you still have it. What may count in Chicago's favour is the complete absence of an engineering department, and concentrations entirely in potential Nobel-winning fields (no so-called 'professional' programs).

I would also venture that institutions that insist on a rigorous core curriculum (in Chicago this is skewed towards the humanities rather than the sciences, but nevertheless) tend to attract students of a certain mindset. It sounds like MIT has a more flexible core than CalTech's and Chicago's old core (the current one is a shade of its former self).

Anonymous said...

I thought Swarthmore is second in the rate of undergraduates going on to win the Nobel Prize.. check out the wikipedia swarthmore page

Anonymous said...

Adding to the previous post:

From wikipedia, I found out that 13 Chicago undergrads won the Nobel, while 5 from Swarthmore won the same prize.. so Chicago's undergrad Nobel laureates is 2.6 times that of Swarthmore.. I cannot find the total number of undergraduate alumni of both Swarthmore and Chicago.. but the average undergraduate class size (according to wikipedia) for Chicago: 1225.25; Swarthmore: 371.. so Chicago's class size is 3.3 times that of Swarthmore.. far higher than 2.6

Anonymous said...

Check Normale di Pisa.

placebo said...

MIT=14th for research impact (citations per faculty member),CALTECH=1st for research impact (citations per faculty member)

MIT=14th for faculty/student ratio, CALTECH= 6th for faculty/student ratio

we have a winner

Blog Archive