Student-to-faculty ratio
(N_S/N_F ratio). The first part, i.e., how many graduate students does
the department have, is fairly easy to compile, either from the
department website (if a department does not have a current list of
their graduate students, that is not a good sign), or from the already
mentioned Roster
of Astronomy Departments (that will usually be one year old). The
second part is harder, as what you actually want are a number of
faculty who can potentially and successfully advise you. So again,
check the department website as the first guess, and count only full
time assistant, associate or full professors. Postdocs, lecturers,
adjunct faculty and other astronomers from nearby or affiliated
institutions might now and then advise a graduate student, and can of
course be excellent advisers, but it is usually not their actual
job. So if a place X says "we have N faculty in our department, but
there are also M PhD astronomers just across the street in the Lab you
could work with", you still pretty much should use N as the upper
limit for N_F. To get a better estimate of the "true" value of N_F,
next go to the ADS Abstract
Search, and for each faculty member see how many papers they wrote
with a graduate student as the first author in the last 5 or 10 years
(for younger faculty you might have to use a smaller number of
years). After removing faculty who do not seem to be working with
graduate students anymore, you now have N_S, the "true" N_F and
therefore N_S/N_F. As you will notice, for most astronomy departments
"true" N_F is somewhere around 15+-2 (one sigma), while N_S varies
from less than 20 to as many as 50, so the student-to-faculty ratio
varies from little more than 1.0 to about 3.0. This obviously is a
huge difference.A somewhat related issue: it is a fairly risky strategy to pick a department because it has one specific faculty member who you absolutely have to work with. First of all, check using ADS if that person is still scientifically active. Second, ask how many graduate students that faculty already has and if he/she could take you on too. Third, faculty might move (I have seen some departmental websites with faculty names that were no longer there). Finally, what if you find out after a month (or worse, a year) that you absolutely can't stand that person, despite their international fame and membership in many prestigious committees and organizations? Can you then switch to a different adviser and research topic without any problem?
Median and mean years to
graduation over the last decade. Typically, that should be 5-6 years,
and in vast majority of cases there is really no reason to be an
astronomy graduate student for longer than 6 years.
Graduation rate, simply
defined as: the number of people who left with a Ph.D. divided by the
number of all students who entered the program. This is harder to
obtain by yourself, as the names of these who dropped out are usually
not listed under "See how great we are" banner on the departmental
website. So simply ask them, they should know, and if they do not know
or don't want to tell you, that again is not a good sign.
Fortunately, in that case you can use again the Roster
of Astronomy Departments and an approximate assumption of a steady
state to arrive at your own estimate: you simply add all the incoming
first year graduate students and all the awarded degrees over the last
7-8 years, and if some department during that time had admitted 50
first-year graduate students, but only awarded 30 PhDs, their
graduation rate is about 60% (by the way, that seems to be the
national average). I have run the numbers for our department and
getting the estimate this way actually works very well.
Next number: job placement
rate, defined as the number of those who took a job in astronomy after
graduation, divided by the total number of those who graduated with a
PhD, summed over a decade or so to reduce Poisson noise. Here you
have to rely on the department providing you with these data, but some
of them (including
ours) keep an updated WWW list of their past PhDs (and again, I
would really wonder about those who don't). This is important
because, as I stated already, I assume you want to go to a graduate
program in astronomy to become a professional astronomer, and not to
use your astronomy PhD to make money for a hedge fund (assuming some
of them will survive that long). At this point you could multiply
these last two ratios to arrive at your Bayesian likelihood of still
working in astronomy 5-7 years from now, if you decide to join given
department.Note that many departments (including ours) will brag about all these Hubble Fellows and other prize postdoctoral fellows they have produced over the years. That is obviously a good thing, but almost anybody can "turn gold into gold", i.e., many of these people would probably succeed no matter where they went (unfortunately, attempts to view their career paths in "parallel universes" have not been successful so far). Beware however of the departments which routinely "turn gold into lead": you don't want to undergo that particular transmutation, as it is very hard to reverse. I am not trying to say that there are astronomy departments out there that on purpose set out to impede their own students' careers. That would be a crazy thing to do, and most astronomers are not crazy. But some departments, as long as they have "sufficient" number of successful graduate students, seem to be less focused on what happens to the less successful ones, which might explain their lower graduation rate and lower job placement rate. To use baseball analogy again, these departments are like some baseball clubs which, despite having some highly visible and highly paid star players, consistently fail to make playoffs (no, not a perfect analogy).
Another important data
point is the publication rate of graduate students. Ask how many
papers (mean and median) do graduate students produce in their first
two years (you probably want to start doing science as soon as
possible) and how many by their graduation. Publishing papers is the
only true "currency" in astronomy (and science in general), and in
astronomy especially valuable are first-author papers (see this paper for more
discussion).
Ask about how the financial
support is organized: will the person you work with be paying you, or
is there a way to work with a faculty member even if they don't have a
grant to study this specific problem (I call that "scientific
freedom")? If you will be supported as a TA, ask how many hours per
week that typically involves (remember, we assume your goal is to do
science); you will want later to verify that number during your visit
by asking graduate students. Ask about cost of health and dental
insurance (and if you have family, cost of health insurance for your
spouse and/or kids). The amount of the actual stipend is also
important, but after adjusting for the cost
of living everybody pays more or less the same (but check in each
case, nobody gets rich while being a graduate student, but you want to
have some money left each month after paying for food, rent, etc.,
etc.)
November 2008