Mapping Academia

Even though individuals display strong correlations between their verbal/writing and quantitive GRE scores, Razib Khan observes that the average GRE scores by intended major show little correlation. Khan also notes:

Philosophers are the smartest humanists, physicists the smartest scientists, economists the smartest social scientists.

I wonder: why do these also happen to be three of my four favorite academic areas (the other being computer science)? Could some areas be better suited to high IQ folks?  If so, am I attracted to those because I think I’m smart? This conflicts with my impression that I like these subjects because they seem objectively more interesting, but that could just be my rationalization.

44 months ago I posted on an interesting “map of science,” and digging deeper today I find that in ’09 the folks who made that map merged twenty different maps of science/academia into this 2D consensus map:


It seems that academic fields naturally form something like a circle, with no fields being especially central. Especially interesting to me, the fields I prefer are all clustered together on one side; my history was to move from E to P to H to CS to an M-style SS.  These topic areas seem to roughly have higher GRE scores, to involve more general and abstract reasoning, and to discuss “far” things further in space, time, social distance, and hypothetically. Apparently academia is divided by near vs. far topics, with math and IQ more important for the far topics, even though math and other formal analysis invokes a near mental mode. The axis orthogonal to near vs far seems to be living vs. dead. Why does academia distribute itself as a circle in this two-dimensional space?

Added 9p: The MapOfScience website, where I got that ’07 graph I liked, now only offers this one:


If you don’t look carefully you won’t notice that the right and left sides actually connect.  Apparently the idea that social science is closely related computer science offends folks there, just as it seems to offend 3 of the 4 comments here so far. More hating on econ?

Many point out that this SS-CS connection seems one of the weakest links in the consensus ring, but that is in part due to the fact that the databases used to generate these maps usually only include data on “sciences”, from which the humanities and many social sciences are purposely excluded.  There has long been a campaign to marginalize these areas from the main body of academia.

GD Star Rating
Tagged as: , , ,
Trackback URL:
  • ZachK

    for those of us who are, professionally, in Bio, but are hobbyists in “high IQ areas” (yes, talking about myself) how can we signal to others, that we are high IQ.

    My attempts: learning math/CS, getting into areas which are computational, modeled using mathematics, use ‘on the margin’ thinking (ie- evolution, ecology)…

  • darghe

    Kinda disappointing that one weak link between CS and the social “sciences” keeps this graph from being a straight up spectrum of hard to soft science.

    • Doug

      I work in a field that straddles the social science/computer science connection (quantitative finance). I believe the connection is very real, and not just an aberration. I wish they would have included nodes encompassing data mining, machine learning, statistics.

      I believe then the connection between CS/SS would have looked much stronger. Sort of like the same bridge math is from CS to physics, machine learnings is from CS to social science.

    • Buck Farmer

      Arguably there’s a pretty big divide within SS…

      …I know many journals in Econ that probably have never cited anything in another SS let alone the Humanities. Citations of mathematics and statistics are far more common than of cultural anthropology, X studies, etc.

      I suspect you’d see a similar lack of cross-citation if you looked at a journal of X studies.

  • following darghe, i wouldn’t put too much stock in this being a ring as a single weak link can turn a chain into a ring or a ring into a figure eight. you can see this in this Science article or run a simulation of it in NetLogo (requires Java VM)

  • dWj

    I agree with the previous points. What you have is mostly a line, from physics up to systems of physical particles up to systems of systems, etc., with the top and bottom linked by a methodology. It’s a little bit interesting, though, that the methodology doesn’t find as much use in between. (I would have expected a link between the math/CS cluster and infectious disease — can we have an epidemiology node?)

    • MikeL

      The CS-SS link is mostly a one way link of SS people needing CS people to do their work, but CS people don’t really need SS people much, other than a few areas like social networking.

      • Eric

        Well, there are issues like:

        -economics of information
        -user experience
        -law and privacy
        -information and organizations
        -concept modeling, knowledge management (also Semantic Web / linked data)
        -information systems architectures, design

        All of these have much more “two-way” collaborative flows than simply social scientists leaching off of computer sciences. In these issues social science methods and perspectives are equally important.

    • Zach K

      I too was thinking that this graph was missing some nodes between math and infectious disease epidemiology and ecology.

  • The distribution of those actually admitted is probably much different. I recall the distribution of scores on the Miller’s Analogies Test, which correlates strongly with Verbal GRE. The averages the tester, the Psychological Corporation, supplied were based on admitted first-year graduate students at selective institutions. At the very pinnacle, by a large margin, was physics. Remember, we’re talking about verbal comprehension and reasoning here, not quantitative ability. The second, well behind physics but substantially higher than the pack, was psychology. At the bottom, as you would more or less expect, was education. Social work was also low. The humanities (not differentiated) were combined, and the score fell in the middle. Chemistry was a lot lower than physics.

    This was decades ago (funny what you remember). But in evaluating how challenging the discipline, it might matter a lot whether you’re talking about undergraduate majors or graduate admittees, raising the question of which you should consider when inferring how difficult or abstract a subject.

  • David

    A guess about why smart people cluster to certain subfields: I suspect they’re attracted by the higher research standards in a that field, while people who have trouble living up to those standards are intimidated into finding something else interesting. It works like this within the disciplines. In my field, most of us aren’t smart enough to contribute in the philosophy of physics, so we settle for metaphysics. Failing that, we slum in epistemology, failing that, in applied ethics. This hierarchy generally undisputed, though it is universally remarked that the value of the contributions in the various fields does not correlate with the difficulty of the fields. I’d even go so far to say that the philosophers of physics are not working on many inherently important problems.

    Empirical evidence for this hierarchy is this: When a philosopher of physics decides to toss off an article in one of the “lower” fields like metaphysics, it is often significant contribution. The converse doesn’t hold: metaphysicians don’t make significant contributions to the philosophy of physics. However, much recent progress in epistemology came from metaphysicians that moonlighted as epistemologists. Epistemologists are pretty hopeless on the turf of the metaphysicians, though. There are enough such asymmetries that you can make a very reliable map. I wonder if other academic fields have something similar. I want to see more maps!

  • Michael Lin

    I’ve long thought that economics and computer science are appealling in a similar way, with economics falling on the slightly more human side, and computer science falling on the slightly less human side.

  • bh

    someone has to link to it…

  • Buck Farmer

    Is north-south axis fundamentally about deductive vs. inductive approaches, methodology vs. subject-focus…I feel there’s a divide between knowledge-acquisition, direct application, and social/policy change, but it doesn’t seem to fit into the current 2-dimensional mapping.

    My personal path has been from Math -> Physics -> Economics…then I graduated, though if I’d been around two more years I might have moved on to Philosophy. I also dallied with Linguistics briefly.

    For me, methodology was key. If the methodology was sound I had no qualms about applying it to any problem or field that passed my interest.

    An abiding argument I’ve had with my friends who concentrated in English is over what English’s methodology is, what are its criteria for good research, etc.

    I’ve also been horrorified by friends in the sciences who don’t know their Popper and Kuhn.

  • DK

    Why does academia distribute itself as a circle in this two-dimensional space?

    Duh. Because when everything is related, a circle can always be constructed. White House, poisonous lizards, omelet and MS Word are parts of a circle, too. Nothing’s special about academia.

    • You must indicate the respect in which the items are related, as here but not in your example. Everything may be related to everything, but not in every respect.

    • And because they left out a lot of minor fields that would create cross-links: human factors between engineering and psychology, mathematical geography/cartography between math and earth sciences, and bioinformatics between CS and biology for three quick examples.

  • What I am really surprised about is the lack of links between HS and I, HS and B, N and B, etc.

    • Eric

      Humanities are now seeing rapidly increasing links with Computer Science. The field of “digital humanities” is growing fast.

  • candy

    No link between neuroscience and computer science?

  • JS Allen

    The most interesting fields will have the highest number of applicants, and will thus be able to be more selective. So it makes sense that the most interesting fields will have people with higher GRE scores.

    You’ll see the same thing in the military. IIRC, infantry is the highest demand (!) in army, so for the slots that are awarded purely on merit, you’ll have the highest IQ members in infantry.

  • michael vassar

    Simple model. Every field is dominated by people about 2 standard deviations smarter than its least intelligent members. Anyone beyond that is likely to have standards for work high enough to antagonize its median member. The fields where the highest levels of intelligence are required to participate at all will therefore be dominated by smarter people (even if in absolute numbers there are more people of any given level of intelligence in larger fields like biology and psychology). The dominant figures in a field determine what that field’s participants spend their time reading and what experiments they do, and thus the attractiveness of the field to smart people.

    All of academia is far mode. History is about as near-mode as academia gets and about as far mode as normal people get.

    BTW It seems obvious to me that the social sciences and humanities are not marginalized.

    • Jess Riedel

      BTW It seems obvious to me that the social sciences and humanities are not marginalized.

      Could you elaborate?

      Due to the nature of funding, the experimental sciences get much more public money and, therefore, attention from the university. Graduate students in the hard science generally get their education for free (either through fellowships or greatly subsidized TA-ships) but the humanities students pay tens of thousands of dollars, which doubtlessly lowers the quality of both the graduates and the research students conduct on behalf of their professors.

  • I think there’s only one main dimension. Simple vs. Complex. tracked by Near vs. Far respectively. Since simple is amenable to logic chopping, and complex best handled by feature detection. Note that this is wrt. methodology, not subject matter. Dead vs. living also correlates because the dead tends to be simple, and the living complex.

  • Reminds me of the hierarchy of the sciences, whose position predicts the frequency with which negative results are published.

  • Hey, what about us humanities folk who use evolutionary biology? Yes, we typically go through evolutionary psychology to get there, but some of us use biology. More, those who look for patterns in texts have to use computer science. Where’s the link there? And there a few of us who use fractal geometry too. I know there are only a few of us, but shouldn’t that get us at least a dim line?

    • William H. Stoddard

      I think you guys are in the minority. Some years ago I visited an old friend in Amherst for her fiftieth birthday, and she took me on a tour of used bookstores, knowing my tastes. I was really struck by one of them that had an entire wall of literary criticism—because about a third were Marxist and another third were psychoanalytic, at a ballpark guess. I remember feeling a moment of nausea, and thinking, “Don’t you people realize that neither of those theories has had the slightest claim to be valid description of human beings or human societies for at least half a century?”

  • William H. Stoddard

    It would be more useful if we had a chart that treated at least economics separately from the other social sciences, which share its statistical methods but don’t seem to have converged as strongly on theoretical models; and that treated philosophy and linguistics separately from the other “humanities” fields.

    I’m not seeing how you define the two orthogonal axes. The first is philosophy/economics/computer science/mathematics/physics versus what? And what are the opposite poles of the second? I’m not sure how to interpret “living vs. dead.”

  • Pingback: tdaxp, Ph.D. » Blog Archive » Mapping Science()

  • Jordan

    Perhaps the center of the ring is the all-encompassing “science of everything” that we are too stupid to wrap our minds around.