Thursday 4 April 2013

Harvard is a second rate research university


As a companion to my posting on Harvard Medical School, it may be worth mentioning that it is a fact that Harvard is a second rate research university - and I have the numbers to prove it!

Of course, being second rate is still a big deal - but it is contrary to what seems to be the near universal belief of Americans.

First, I should emphasize, that the only remotiely objective evaluations of research quality relate to science - therefore, when adopting an international perspective, the humanities, social science and law are simply elminated from the analysis.  Harvard may excel at these other things, or it may not, but nobody really knows.


What Harvard is, in terms of science is a very high volume research university

Look at Table 2, citations - Harvard has more citations per unit time in the science research literature than any other university.

1 Harvard

2 Johns Hopkins

3 Stanford

4 U of Washington, Seattle


6 U Michigan, Ann Arbor


8 U Pennsylvania

9 U California, San Diego

10 U California, Berkeley


But you can see from the other Universities in this list, that citation volume is not a reliable guide to elite science - and is mostly a product of a university being very big - that is employing very large numbers of highly productive researchers.

When it comes to being excellent at the highest level of scientific research, Harvard is in the second rank and the premier university is without question MIT (Massachusetts Institute of Technology).


For example, analysis using a combined metric of Nobel prizes, Fields medals, Lasker awards and Turing awards for 1987-2006, we get:

MIT 13

Stanford University 10

Princeton University 10

Chicago University 8

University of California, Berkeley 7

Columbia University 7

Harvard University 6

CalTech 5

UCSF (University of California San Fransico) 5

Cornell University 4


What is more, the trend for Harvard is downgoing - because it used to be the best (and MIT was not so special) - Looking just at Nobel prizes looking at 20 years segments from 1947:

Number of United States Nobel laureates by institution – 20 year segments from 1947 to 2006

Institution 1947–66; 1967–86; 1987–2006

Harvard University 9 - 13 - 5

University of California Berkeley 7 - 3 - 4

Stanford University 4 - 5 - 9

CalTech 4 - 4 - 5

Columbia University 4 - 1 - 7

Rockefeller Institute & University 3 - 6 - 3

Chicago University 2 - 4 - 7

Princeton University 1 - 2 - 6

MIT 1 - 5 - 11

Cornell University 1 - 4 - 2


Thus Harvard went from 9 Nobels via 13 down to 5, whereas MIT went up from 1 via 5 to 11.


Thus from the perspective of scientometrics Harvard looks like it is nowadays and increasingly geared-up for hiring highly productive but 'safe' researchers in very large numbers (what Kuhn termed 'normal science'); while the smaller and more truly research-elite modern Universities are doing a better job at recruiting more original/ creative 'revolutionary science' researchers.


[Note - in the above Nobel science prize analyses I included Economics, which I would not nowadays do; but since I discovered that nobody takes any notice whatsoever of this kind of evidence, I find that I cannot be bothered to re-do the numbers.]


dearieme said...

Long ago when I was an undergraduate, all us science-and-engineering types had heard of MIT, but knew of Harvard only in terms of its Business School. How farsighted of us.

sykes.1 said...

Harvard's basic problem is that its faculty is superannuated and corrupt, plagiarism being the main sin but outright lying fairly common. Think Tribe, Goodwin, Gould, Lewontin, et al.

Bruce Charlton said...

@sykes - Yes, corrupt - but clearly *not* superannuated in the sense of faculty being lazy or unproductive - but I don't think we have any comparative data about whether Harvard is *more* corrupt than elsewhere.

On the other hand, I would agree in the sense that corruption at Harvard is much more damaging than elsewhere, especially in the public sphere/media where they assume that the motto of Veritas remains correct - whereas I would hazard that it ceased to be correct around about when they quietly deleted "Christo et Ecclesiae" (for Christ and the Church) which for a long time used to surround the heraldic shield.

Marcus B. Fairchild said...

Professor Charlton,

Your assessment assumes as obvious that NLG prizes are the proper criteria for revolutionary science and scientific excellence. But why should that necessarily be taken for granted? You've made your anti-democratic views quite clear in the past, and I tend to agree. But aren't the NLG prizes a form of democratic elections, albeit with an electorate of experts?

As for Harvard's size, have you looked into Harvard's citations or publications per faculty member? Harvard is big, but it's smaller than the U. of Michigan and Berkeley, for example. What about their NLG per faculty member, relative to Harvard's?

MC said...

Don't scientific researchers lean heavily on their students to do the legwork in their research, and perhaps to spur their creativity? I don't know since I majored in history.

Harvard is the finishing school for the American elite, so its students are going to be less help in that regard than the future engineers and scientists at MIT.

Bruce Charlton said...

@MBF - People can always ask for more detail or criticize the variables; but are there any better? These numbers took me and Peter Andras more than 18 months of work (albeit not full time).

If I calculated more or different numbers, would it really make any difference? If people ignore the simple numbers, they'll ignore the complicated numbers.

No, I don't think that Nobels etc are a reliable measure of revolutionary science - they represent what the peer review cartel within science believes is the most important work. If none of the work is important, or it it is all corrupt, then the prizes don't mean much.

All this data says is that by the prevailing criteria of evaluation, Harvard does not rate at the top. But that is interesting, isn't it?

AlexT said...

The smartest and, crucially, most sensible grads i met while living in the US were MIT and CalTech people. They seemed to have quite sensible approach to life in general which differentiated them from the leftist ivy league crowd. Not in all cases perhaps, but in enough to make a difference in my view of them.

Sylvie D. Rousseau said...

What these numbers tell me is how enhanced is applied science nowadays compared to theoretical science 50 years ago. I hazard it might be because practical results are more obviously proof of theoretical accuracy. It is a pity that theoretical science has, for whatever reasons, become less important than applied science.