In case you ever wondered which departments dominate my discipline — anthropology — in America, here we can get a pretty clear sense of demographic dominance, at the very least. I’ve added together the total number of PhDs awarded by each of these departments over the last two decades (1987-2007, 21 years total) and we can see that some departments have produced far more than their share of new doctorate-wielding anthropologists.

UCLA and Berkeley are tied for the greatest production of scholars, at 322 total, with Chicago next at 296, Harvard and Michigan some way behind that at 253, University of Texas-Austin just behind there at 248, University of Florida at 220, University of Arizona at 219, Columbia at 211, and then on down the line.

Not terribly surprisingly, there is a good deal of overlap between this list of the largest graduate departments and the highest ranked departments in 1994’s National Research Council rankings. The orders of the departments don’t correspond perfectly, though; for instance, UCLA was #9 in the rankings but is #1 in terms of its output of scholars.

One of the most striking things about this graph, to my mind, is how it shows that the top 20% of departments have produced a highly disproportionate share of all Ph.Ds. The top producing departments are way ahead of the ones lower down. Consider what happens when we aggregate the graph into fifths (“quintiles”):

What this shows is that the top 20% of departments produce almost half of all doctorates (45%), while the top 40% of departments together produce 67%. This of course isn’t anything like as unequal as the distribution of wealth in America. But it does indicate that the biggest departments, in terms of their production of scholars, yield a very large share of the anthropologists in our profession. The top three departments alone, Berkeley, UCLA and Chicago, produce more than one in ten of all new anthropologists (11%).

I suspect that there are some effects of large scale at work here, such that it becomes easier to maintain a large department once it is already in existence, because it will have a large network of alumni who can then help to find jobs for new students coming from that department, because it will spread its reputation wider and wider across the discipline, because this wide reputation will then in turn be helpful in attracting new students. Correspondingly, on a personal, subjective level, the anthropologists who I know or whose work I know come disproportionately from the bigger departments on this list. Demographic dominance is of course not identical with intellectual dominance in the discipline, but there does seem to be a fairly definite connection between them.

All the while, of course, the tiniest departments are very tiny, producing just a handful of Ph.Ds over the decades. Consider another graph, showing the distribution of departments that produce a given number of doctorates.

This graph has a long tail: a relatively small number of high-producing departments are out there on the right-hand side, isolated from the majority of the departments, which concentrate around the 50-130 doctorates span. (Again, this is a graph of how many departments have produced how many doctorates in total over the last two decades.) I have no idea why there’s such a spike at the 100 PhDs mark, but that’s clearly the modal point in the graph. Curiously, as we go off to the left (and see also the bottom of the first graph above), we find a bunch of departments that have produced virtually no PhDs at all: Alabama, Arkansas, Maine and Wyoming have each produced one single doctorate over this period, while Nevada has produced three, and Florida State and the California Institute for Integral Studies have each produced nine. Nine PhDs over 21 years is less than one each year, so one has to imagine that there are some departments with very tiny PhD programs (probably they have larger MA programs with just the occasional doctorate conferred).

I have the feeling looking at these graphs that demographic dominance is one of those brute facts that probably exerts its effects in silence, by dominating not so much any particular outcome but rather the space of possibilities, of ideas, of available people that constitutes the discipline’s historical trajectory. It does make a difference that right now, today, there are 322 Berkeley grads floating around out there in our field, whereas there are only 155 from CUNY and 71 from Minnesota and 36 from Utah and 14 from the University of Illinois-Chicago, six blocks from where I lived last year. Remember, this is 322 Berkeley grads over the last two decades: they don’t constitute a single social group, mostly likely, but they probably do have some commonalities that are products of their departmental and institutional origins, they do constitute a large realm of recognition as against a smaller department’s small realm of recognition.

What kind of difference this makes is not so easy to pin down. But I do recall that when I was in college and I proposed going to Santa Cruz (which has produced 38 PhDs in 1987-2007, i.e. less than two yearly), my college anthropology teachers said: I’m sure it’s great, but it’s tiny and you’d never get a job. I don’t know if that’s true in this case, but one thing seems clear: size matters, when it comes to anthropology departments today. We have, in fact, a somewhat centralized discipline when it comes to the manufacture of scholars. One might reasonably ask whether that is a good thing and what other forms of disciplinary organization might look like or whether they are even imaginable.

But things look a bit different if we turn to look not at undergraduate degrees but at the doctoral degree production that’s essential for the reproduction of the teaching and research body of the profession. (Haven’t had time to look at Master’s degrees so far; I suppose that master’s degrees would serve a joint role as both an intermediate academic credential and a semi-professional credential, and are a stepping stone to the doctorate in some cases, but this requires more research.)

At the doctoral level, anthropology is no longer at the bottom of the charts; over the past forty years it has climbed from being one of the smallest social science graduate fields to being roughly similar to sociology. In 2007, anthro graduated 519 new PhDs while sociology was at 573. Economics, nonetheless, clearly appears to be the dominant social science discipline (demographically speaking), though political science has approached it on several occasions and even surpassed it for a few years earlier this decade.

An interesting thing, in passing, is that some of the small dips and rises in these fields seem to be matched across fields — the little dip around 2002, for example — which suggests that some of these shifts may be results of larger institution-wide changes, and not always discipline-specific; there might be, for example, global reductions in student intake. It certainly sounds plausible.

We do see clearly that the two small, relatively “dominated” fields here are linguistics and ethnic/area studies. They appear not only to be small but also to have stayed relatively static in size over the last few decades — while the other disciplines have seen what looks like more of an upward swing since the 80s. We can see their relatively tiny share even more clearly if we examine a graph of comparative shares of social science doctorates produced over time.

Anthropology, shown at the bottom of the graph, has actually grown in comparative importance since the 60s – it produced 6.3% of social science doctorates in 1966, but is now up to 13.8% (in 2007). Political science and economics loom large in the center of the graph (currently producing a 58.9% majority of all of social science doctorates). It crosses my mind that economics PhDs may have more non-academic job possibilities, meaning that its comparatively larger production of scholars may indicate not a proportionately large academic job market but only that a lot of economics PhDs are getting jobs in government or the private sector. According to CIRGE, who surveyed social science PhDs five years out of grad school, about 18% of social scientists overall worked outside academia; they left economics out of their survey, though, where the percentage may be higher. At my university, 49% of economics PhDs get jobs outside academia, but that may be on the high side of things.

As you can see here as clearly as before, area/ethnic studies and linguistics are small slivers of the total production (at the top of the graph). It’s interesting, however, to compare their situations at the doctoral level with their situations at the bachelor’s level. At the bachelor’s level, linguistics is tiny while anthro and ethnic/area studies are somewhat larger, whereas at the doctoral level, linguistics and area/ethnic studies are both tiny while anthropology remains somewhat larger. In other words, area/ethnic studies is comparatively much larger at the BA level than at the PhD level. Consider the following.

This graph compares the number of BAs to the number of PhDs produced across these three fields. Basically, in comparison to anthropology, it seems that there are way more ethnic/area studies undergrads getting degrees than there are ethnic studies PhDs being produced. My guess is that ethnic/area studies are relatively popular undergraduate fields, but that many people teaching in such fields originally get their PhDs in a more traditional discipline (like anthropology, history, etc), so we see a relative disproportion between the output of scholars (PhDs) and the output of students (BAs).

(I had to plot this on a logarithmic scale because otherwise it would be impossible to show the anthro and ethnic/area studies BAs and PhDs on the same graph; since the PhDs are around an order of magnitude less, they would vanish into the bottom of the scale if it were linear.)

I thought about this last comparison some more and, just for another point of reference, put together one more curious graph, which I’m honestly not quite sure how to interpret yet.

This is a graph of the number of BAs to the number of PhDs for a given year, across fields. OK, so, lower means that fewer BAs are awarded per PhD. In essence, it’s a student:teacher production ratio. It could also, leaving aside inter-field migrations like the (probable) crossover between anthropology and area studies mentioned above, be taken as an index of how research-oriented a given field is; lower means that the field is relatively more oriented towards the production of new scholars, higher means that the field is relatively more oriented towards the production of new bachelor’s holders.

As you see, linguistics is the lowest — they have relatively few undergraduate majors compared to the number of new linguists they produce at the doctoral level. Anthropology is pretty consistently second lowest in this regard. Sociology and political science produce the most BAs per PhD; economics is sort of in the middle, and ethnic/area studies has been climbing, which suggests that their undergrad production is growing at a faster rate than their graduate production.

I will confess, it’s strange to think about these quantitative dimensions of the “production of scholars”; I don’t experience graduate school as being like an assembly line where you come out the end as a figure in a large spreadsheet. Though, of course, that sense is itself revealing of the fact that graduate students are valued for being different from each other, within given disciplinary boundaries. You could say that we’re mass-produced to have fantasies of individuality.

Now, I have to add that at an experiential or practical level, it’s not clear that these disproportions between the number of new economists and the number of new anthropologists make a huge difference. I virtually never interact with economists, so it would seem to make almost no difference how many of them there are; the disciplinary barrier between us is practically impenetrable. Disciplines’ statuses in academia is in some ways more like the US Senate than the House of Representatives: to get by in many contexts, it suffices just to have a discipline, and more power doesn’t necessarily accrue to the largest ones simply by virtue of their size. Still, it’s easy to imagine that institutional resources are distributed in some relation to the size of the disciplines. Larger disciplines = more office space, more faculty salaries, more facilities… not that these resources are always allocated in direct proportion to departmental size: on our campus, they recently announced a $100 million Milton Friedman Institute for Research in Economics, and the very same day they also inaugurated a $0 million dollar Edward Sapir Anthropological Research Center, along with a $0 million dollar Robert Park Sociology Institute and a $0 million dollar… well, you get my drift.

There seem to be complicated, bidirectional relations between the success of a department and its treatment by its administration. Are large departments well funded because they’re large, or are they large because they’re well funded? On our campus, for example, I believe that the size of incoming graduate student classes is determined substantially by higher administrators, who allocate slots for incoming students. They consult with the departments about how to allocate resources, but they also control the money and therefore tend to have the power, in the end. So it seems worthwhile asking ourselves the Marc Bousquet question here: to what extent should we be analyzing the relative size of the disciplines not as an independent phenomenon but as a result of higher layers of administrative decision-making?

Coming up, I’m going to take a closer look at graduate funding and departmental dominance within anthropology itself.

But I didn’t follow up on that at all until I recently started reading the “socio-demographic” work of Charles Soulié, a Bourdieuian French sociologist of universities whose research interests are fairly close to mine. The premise of this research is something like this: by examining the comparative history of enrollments and teaching jobs across disciplines, one can examine what Soulié calls the “evolution of the morphology” of academic fields. This isn’t very hard-core quantitative research by statisticians’ standards, I note — he doesn’t exhibit tedious anxieties about the uncertainties in his sources, nor does he propose mathematical models or major statistical analysis of his data. The methodology seems to be, in essence, visual inspection of the evolving demographics of disciplinary enrollments. He takes these as indicators of things like the “relative position of sociology in the space of disciplines,” and comes up with findings that are like:

• Sociology produced half as many graduates in philosophy in 1973, but now things are reversed, and in 2004 sociology produced 2.6 as many graduates as philosophy. This is an indicator, for Soulié, of sociology’s rising comparative importance in the university system (and philosophy’s stability, which in context was a relative decline).
• In 1998/99, “the fraction of children of professionals and upper management rose to 28.4% in letters and human sciences, against 23.1% in sociology and 38.1% in philosophy” — which tells us something important about the comparative class basis of sociology vs. philosophy at that point in time [updated to clarify: these examples refer to French academia].

I find this kind of thing quite interesting and revealing – hence this series of posts on the demographics of my own discipline – but I wonder about its epistemological basis. What does it mean, actually, that one discipline has more students enrolled than another? Is it right to speak of a competition between disciplines for students? What makes one discipline more “attractive” or “desirable” than another at a given moment? It’s not like students pick their courses based on a completely rational response to a job market, or even an idea market. In fact, it’s not clear that “market” is a good description for these kinds of systems; as Marc Bousquet has often argued, talk about the academic “job market” (for instance) disguises the fact that university administrators actually dictate the academic job system, by deciding to opt for hiring adjuncts, grad students, etc. Likewise, shifts in degrees issued, in enrollments, etc, may not necessarily be the result of “competition” or market forces (whatever one’s stance on the empirical existence of said market forces). There can be other kinds of systematic processes at work; the “morphology” of the disciplines as revealed in their enrollments doesn’t tell you everything about processes of interdisciplinary conflict and coexistence.

But the brute fact remains that there have been major historical shifts in how many students anthropologists educate, and major shifts in how large our discipline is vis-a-vis other disciplines. And these aren’t just arbitrary. They need to be explained, if we’re to understand where our discipline actually exists in the world. When American anthropology is educating a small fraction of a percent of college students, that’s not something that just happens by chance.

I feel here the strong sense of a bias in my own discipline against quantitative analysis. It’s somewhat jarring, from the narrow confines of an anthropologist’s culturalist background, to look at these comparative figures. In cultural anthropology, I think there is a widely shared consensus view today that goes something like this: culture is inherently qualitative, folded over on itself in swathes and patches and wrinkles of rich, dense symbolic significance; it would necessarily be deformed, or at best severely limited, by any effort to reduce it to a general and/or quantitative analysis. Among cultural anthropologists, numbers and quantitative facts are apt to be taken not as means of analysis, but as objects of cultural analysis and symbolic forms in their own right. So we get studies of the cultural effects of perniciously quantifying, rationalizing, neoliberal projects; and we see arguments about how the obsession with the quantitative is itself merely another local cultural phenomenon, and not a privileged, master form of knowing about the world. Often these kinds of arguments are made casually, in passing, or are simply taken for granted, inscribed in our disciplinary habits.

In fact, in cultural anthropology as I know it today, it’s seldom necessary to argue explicitly against quantitative work; its rejection is already inscribed within the positioning of our discipline against its others, like sociology and above all economics. Our primary research method, field ethnography, is almost constitutively anti-quantitative, being oriented overwhelmingly towards the experiential dimension of social life, toward the fine detail of the symbolic, the affective, the discursive. I note that earlier ethnography sometimes involved more quantitative work – Marilyn Strathern’s first book had a lot of figures about pigs in local economies, for instance – and one might speculate that, as post-colonial anthropology lost its identity as the discipline that studied exotic “primitive” cultures, it seized in part on qualitative methods as a new basis for differentiating itself from the other social sciences. And, coming back in the present, I note that even after the fieldwork, later as we ethnographers write our analyses, when we do venture to generalize, we generally make qualitative generalizations (I’m making one right now); and when we do incorporate quantitative information, it often figures as mere background data for our more specific ethnographic arguments. There do appear to be sophisticated studies of mathematical practice, like Helen Verran’s Science and an African Logic which I’d like to read. I note, however, that I’ve seen a number of ethnographic studies of, for example, financial markets, which for all their strengths, still take for granted that their qualitative mode of knowing is drastically epistemologically superior to (and allows privileged ideological diagnosis of) their informants’ quantitative modes of knowing. My sense, in short, is that our rejection of the quantitative can go without saying because a rejection of numbers structures much of our research practice and disciplinary identity. There are, of course, plenty of books about quantitative methods in anthropology, and there are people who polemicize against cultural anthropology’s “rejection of science” and numbers — here’s a pretty loquacious example — but, in relation to the literature I read and the cultural anthropologists I know, these are the exceptions that prove the rule.

Now, the anthropological objection to quantitative research is correct and good up to a point. Numbers certainly are cultural forms, a fact which I have yet to see acknowledged by a quantitative researcher. Quantitative data is of course not always useful or accurate, and can be grossly misleading, can even be a vehicle for various other political and ideological projects. And any social science that effaces the lived experience of cultural life is bound to give a rather reductive analysis of the world.

But it still feels wrong and simplistic to reject and avoid numbers in cultural anthropology in the ways that we often do. Yes, numbers are potentially oversimplifying abstractions, and are sometimes used in ideologically pernicious and theoretically problematic ways. But I wonder what ways of knowing we numb ourselves to, in shrouding ourselves in qualitative, sensuous, tactile accounts of the world. Ethnographic tactility and sensitivity can become its own form of anaesthesia, can’t it?

As with the more general university situation, all fields have been growing, albeit with a major dip in the mid-seventies to late-eighties, which is again probably due to the Baby Boom ending. It’s obvious that the biggest field by far is political science — though my figures for political science also include public administration, whose more marketable vocational potential may explain the overall predominance of this discipline. Economics and sociology, in blue and green, have been somewhat similar for decades — while sociology was far more popular from the ’60s into the ’70s, economics overtook it between 1980 and 1994, and since then sociology has pulled ahead slightly but not that much. One notices a curious correlation, probably spurious I suppose, between the economics degrees issued and the political party holding the presidency: throughout the Reagan/Bush 1 era, economics is ascending; then it drops substantially under Clinton; then it rises again around when Bush 2 comes into office.

I note in passing that linguistics is absolutely tiny and barely visible (a thin brown line at the bottom of the graph). Our own discipline, anthropology, is pretty low on the charts too; and it also has a very close partner on the graph, which is area and ethnic studies. It turns out, somewhat unexpectedly, that anthropology and ethnic/area studies have been very closely linked in undergraduate enrollments since the 60s. Let’s look at this in more detail.

While larger fields like sociology and economics award tens of thousands of bachelor’s degrees yearly, anthropology and ethnic studies have been climbing up slowly to 7-8,000 degrees awarded per year. One can only speculate that the sort of undergraduates who are interested in anthropology are about equally likely to be interested in area or ethnic studies, which would fit, of course, with the fact that anthropology has long been specialized in teaching about ethnic or foreign Others, and overlaps frequently in its content with the various * Studies departments. I do note one interesting difference, though, which is that the usual dip in degrees awarded (that big local decline starting in 1975) is far more pronounced in anthropology, while ethnic studies fell less far. Might it be the case that ethnic and area studies were gaining in comparative importance in that period because of the growing institutionalization since the 70s of multiculturalist education and programs? What can we learn about the history of the disciplines from these analyses?

When we look at the production of doctorates, things look somewhat different than at the bachelor’s level, but I want to stop here for now to think about the methodological implications of this kind of analysis.

I often feel that my discipline, anthropology, doesn’t sufficiently discuss its own structural situation in the academic world. Where do we fit in the ecology of disciplines? In the national competition for student enrollments? How many anthropologists are there, exactly? And what is the structure of our academic labor system; what fraction are tenured, tenure-track, contract, part-time? How many of us work outside the academy? Which departments are dominant or central in our profession?

This thought had been rolling about in my head when all of a sudden yesterday I discovered this tremendously neat website: WebCASPAR, the NSF’s Integrated Science and Engineering Resources Data System. Strange name, I grant you. And at first sight it’s confusing for those of us with basically no statistical background. (Which, for the record, I consider an embarrassing absence from our professional training in cultural anthropology. On which more later.) It starts out with a “Table Builder”; you pick a data source; you pick an “analysis variable” (which is what the content of your table will be full of: numbers of students enrolled, for example); you pick a “classification variable” (which forms one dimension or the other of your table); you set some details about what gets displayed where. And then magically…

…or in my case, after a few hours of experimentation, we have a series of tables: the growing US student enrollments since the 60s, the changing ethnic composition of the US student body, the fraction of US students at public vs. private schools… and guess what? I’m happy to announce that we still have a predominantly public higher education system! 74% of college students are currently in public institutions, it appears. Of course, the survey doesn’t take into account that many public universities are now only precariously supported by public money, and are increasingly selective, more like semi-public, but we’ll let that go for now. Let’s mention only in passing the striking comparative fact that in France, “public” universities are, by law, open admissions and almost free of tuition, or in other words, they’re far more open to the public than their “public” U.S. counterparts (a fact which should provoke more thought than it does).

Coming back to the structural situation of anthropology, I think we have to start by putting it in the historical context of its relation to other disciplines. There are a number of ways of measuring these interdisciplinary comparisons – by enrollment, by degrees granted, by faculty jobs, by funds allotted, by publications produced – which, though they don’t begin to exhaust what might be said qualitatively about anthropology’s place in academy or the broader world, can still teach us some things that might fail to appear from purely qualitative examination or personal experience.

Take the graph we started with above. See that little yellow line almost at the bottom? Social sciences. That’s us. We’re in there somewhere. As I’ll show later in more detail, the last few years anthropology granted about 5% of all social science degrees. And social science in general is currently about tied with humanities for the least number of graduates of any major sector of higher education. As far back as this graph goes, social sciences always been toward the bottom, though it’s interesting to see that humanities (green) used to be larger than us, but fell below in the mid-70s and has closely paralleled our graduation numbers ever since.

In addition to the relative smallness of social science in general, the other major thing to learn here is the extreme predominance of vocational subjects in American higher education. So much for people who fantasize that American higher education is (currently) about liberal arts! They should look at where the students actually are. Business has been the most popular single subject in our country since 1974, surpassed here only by a miscellaneous category of professional and vocational degrees (which includes communications, law, social services, librarians, and other unspecified fields). Education and engineering are fairly large also. Sciences appear to be large, but the sciences category actually includes all the vocationally oriented health fields (nursing, medicine, etc) as well as computer science, and so the pure research sciences are probably much smaller than they appear.

Worth noting, however, that enrollments as a whole have been growing steadily almost throughout this whole period, albeit with a pronounced dip in the mid-80s that registers in the graph above in all fields besides business, engineering and sciences. Frankly I am not completely sure of the explanation for this dip in the Reagan era — there may be institutional and political factors, but the most likely thing I think is that the baby boomers were done going to college and the birth rate fell after about 1963, so there would naturally have been lower college enrollment 18 years hence.

So what does all this tell us? The number of people becoming social scientists is still rising; anthropology is a mass phenomenon, in that it produces thousands of graduates yearly; but anthropology is still only 5% of social science degrees, and social science is only about 5.6% of total degrees granted, out of about 6% of the total U.S. population enrolled in universities… you get my drift? Anthropology granted 8,086 degrees in 2006, out of 197,595 in social science, and out of 3,519,259 in all fields, which means, in other words, that we produce about a fifth of one percent of all American university graduates. So we can see that anthropology is really not attracting a terribly high fraction of American university students, especially when compared to the vocational fields.

This probably means that anyone who tells you about how flexible an anthropology degree is, how much it prepares you for a whole range of careers, is probably indulging in a bit of wishful thinking, because surely if this were the case, all the professional school people would jump ship? And who are these people, this fraction of a percent of students, who enter our discipline? Where do they come from and what kinds of incentives, cultural contexts, social norms guide their disciplinary choice? It would be good to find some figures on the class backgrounds of people getting degrees in anthropology… I’ll keep my eyes open. In the near term, look out for more of these graphs of anthropology’s place in the world.