<p>The PhD rankings are a measure of a lot of things:</p>
<p>a) type of student body, in terms of percentage interested in hard-core academics. No large public university is going to ever do well on a per capita index because the bulk of their students aren't really hand selected for, or have any particular aptitude for, academics. It's like comparing college rates between a private prep school and a public high school. </p>
<p>b) type of career path, PhDs. imply reasearch, academic, teaching, type jobs. </p>
<p>c) quality of the program, which includes research an/or teaching measures. A program can be excellent because it has a lot of reknowned researchers. Or, it can be excellent because of the quality of the teaching, professor/student mentoring, etc. These lists consistently mimic the "best" colleges and universities. In this case, note how closely the raw numbers and/or the per capita PhD. numbers echo Alexandre's recommendations for great schools in his field.</p>
<p>d) Size of the department at a particular school. A useful measure. Obviously, if you want a school with a "strong economics department", you expect that a sizeable percentage of students enjoy the department sufficiently to become majors.</p>
<p>e) Overall, "feel" of the university. As I look at these charts, they remain remarkable consistent across various departments. To me, they indicate a school that, as a whole, is academically focused. Conversely, these lists also tend to be revealing about some "well-known" private universities that aren't all that academically inclined as well -- particularly when you see the same rankings across many departments. </p>
<p>Would I choose a college based solely on any of these lists? Heck, no. Nor would I choose a college based on the "strength" of a single department -- unless it were in a niche field where the typical school simply doesn't offer much.</p>
<p>Like any other measure (SATs, graduation rates, binge drinking rates, percentage of varsity athletes, percentage of frats, diversity, percentage on financial aid, etc.), it's just a data point. It takes many data points to get a sense of an institution.</p>