One of the nice features of the PGR data is the duality in the relationship between departments and specialties. Departmental identities are defined in part by the kind of specialized work that gets done in them. The identity of areas is associated with particular departments and schools (with a large or small ’s'). The PGR data lets us see some of this association, and of course also make the link between this relationship and overall status. Like departments, some areas are judged more important than others.
One difficulty with visualizing the connection between specialization and departments is that there are too many dimensions of specialization, and a lot of departments, as well. On the one hand we’d like a nice picture with a lot of useful information, but on the other hand we need to make it visually comprehensible and true to the data. Here’s one way to do this. It doesn’t succeed perfectly in these aims, but I like it anyway.
I want to get at the idea that departments can be thought of as variable amalgams of specalist expertise. The problem is that the 2006 data has too many specialty areas to easily visualize. So I combine a number of areas that the data says are very highly correlated within departments: 17th and 18th century philosophy are grouped together; medieval philosophy and philosophy of religion; ethics and metaethics; general philosophy of science and the special sciences; 19th and 20th century continental philosophy; and philosophy of math, logic, and decision theory. Unfortunately I also have to drop some specialty areas, based mostly on the small number of programs rated in those areas. (I said this was going to be not quite satisfactory.) That gets us from more than twenty nine areas down to twelve, which is just barely manageable.
Our twelve specialty areas contribute to a department’s identity or profile. Imagine each one of them is a segment or wedge in a circle, a bit like an old Trivial Pursuit gamepiece, but with twelve wedges instead of six. Yes, I know that the perceptual qualities of wedges are not good, but bear with me. Suppose a department that is top-ranked in every single one of these specialty areas looks like this:
Now we scale the wedges so that each segment is proportional to the department’s rank in that area. The smaller the wedge, the lower the reputation. If we order departments by their PGR ranking, here’s what we get:
A lot of information is summarized in this figure. What does it tell us? In 2006, Oxford most closely approximated the key—i.e. it came closest to the top in the most subfields—but it was not the top-ranked department. (It was the largest department, in terms of faculty, by some distance.) It’s clear that not all specialty areas count equally for overall reputation. In 2006, NYU and Rutgers were weak or had very little reputation to speak of in a couple of areas, but still outranked Oxford. Similarly, the other top departments all have gaps in their coverage. At the same time, top departments are good at a lot of things. It’s not enough to specialize in just a few areas. Amongst the top twenty departments in 2006, MIT and the ANU had the narrowest range, relatively speaking, but their strength was concentrated in areas that are very strongly associated with overall reputation—in particular, Metaphysics, Epistemology, Language, and Philosophy of Mind.
The further down we go, the more variegated departments begin to look. The specialty segments in the plot are not randomly organized: the top half, roughly, are contemporary specializations and the bottom half, roughtly, are historical specializations. More importantly, reading clockwise from the M&E segment they are arrayed in order of how well scores in each specialty area correlate with the overall PGR score for departments.
Other things to notice: there are only a few really specialized departments, with a lot of strength in just one or maybe two areas. The LSE (in philosophy of science) is one example. Some departments look like mirror images of each other, especially through the horizonal axis of the circle. If you want to get a sense of returns to specializtion, compare MIT with Penn, for example, or Sheffield with McGill. And you should also note that the figure clearly doesn’t capture everything relevant about the overall reputation of departments, especially ones further down the rank order: there are several cases where adjacent departments have very different profiles, which suggests that raters are assessing differences between them in ways the figure doesn’t capture.
Finally, it’s interesting to pick out departments may be relatively far apart from one another in the rankings but which seem to be the same sort of departments—they have the same pattern of specialties, if not the same reputation in those areas. In my next post (I am rapidly running out of time here!) I’ll present a way to visualize that idea.