Comps

February 3, 2008 9:32 AM

I have comprehensive exams in 230 days. Any studying advice?

In 230 days I am taking my comprehensive exams for my political science PhD program. These tests determine whether I continue on to write my dissertation or (if I fail twice) get removed from the PhD program.

Briefly, the exam is in two parts. First, I receive a set of questions on international relations and have to write a 50+ page answer in 48 hours. Second, a week later I receive a set of questions on quantitative research methodology (game theory, statistics, formal modeling design, etc...) and I have to write an answer in 48 hours. The entire test is written, no oral examinations. In plain English, the first exam tests my knowledge of other people's research, the second exam tests my knowledge of programming and mathematics.

I have classes and work for the next 130 days, then 100 days of summer with no classes or work. I have enough money saved so that I do not need to earn an income over the summer. I do not thrive on stress (like some people) and therefore want to start studying now rather than wait until the end.

I know many of the Hivemind have taken comprehensive exams or other similarly large tests (USMLE for example).

1. How should I go about studying systematically?
2. Do you have any tips regarding studying for large tests?
3. I have to use R (a programming language) for the statistics, how can I go about learning R well?
4. How should I organize my notes?
5. Any general advice?

Thanks so much for any replies!

posted by chrisalbon to work & money (11 comments total)

how can I go about learning R well?

I highly recommend An Introduction to S and S-Plus by Phil Spector. It is short and easy to understand. I advise against attempting to learn from the introduction provided by the R Core Team.
posted by grouse at 9:38 AM on February 3, 2008

When I was studying for my comprehensives, I made a detailed reading calendar--right down to the hour--and stuck to it. If that meant studying while eating lunch or on the #6 bus to downtown Chicago, then so be it.

However, I also scheduled in relaxation time. Allow yourself time to do something mindless, or the words will start doing tap dances on the page.

Part of my reading calendar included typing up detailed notes after I finished each text; as I got closer to the exam, I spent more time reading over the notes and highlighting the relationships between texts.

How much guidance will you receive ahead of time? (If you haven't already done so, you should be looking at the scholarship of your various examiners...)
posted by thomas j wise at 10:27 AM on February 3, 2008

I took notes on everything I read, and then made huge flowcharts on butcher paper in an attempt to make a visual representation of how everything fit together. I am a particularly visual sort of person, so I think the flowchart was especially useful for me — it forced me to really remember what I'd learned when I had to think about where it fit into The Big Chart.

My examination was oral, but I think my note-keeping system would work for you as well, especially if it's the sort of thing you'll be able to refer to during your exam.

Is your exam administered by your Ph.D. committee? Or any committee at all? If so, the advice from thomas j wise about understanding their own research is invaluable. I managed to flounder through a few questions only because I could compare what I was doing to the work of my examiners.
posted by adiabat at 10:38 AM on February 3, 2008

I learn which professors are writing questions for the exam six weeks ahead of time.
posted by chrisalbon at 10:51 AM on February 3, 2008

Concentrate your revision into some smaller key areas that you know will be covered, divide things up into smaller pieces of information. Just to make it less daunting.
Sounds hardcore and Im writing my thesis at the moment, good luck to you!
posted by Neonshock at 11:14 AM on February 3, 2008

I would recomend using Gelman and Hill's
Data Analysis Using Regression and Multilevel/Hierarchical Models to learn R. The book is very clear and surprisingly cheap. I also think it might be a great fit for you because Gelman is one of the best statisticians working in political science. Many of the examples are focused on political science questions.

As for the comps: the biggest challenge is to create a schedule and stick with it by doing a little bit each day. The more you make studying routine, the easier the whole process will be. I wish I had been better at managing my time when I was doing my comps.

Good luck!
posted by eisenkr at 12:04 PM on February 3, 2008

As you build your schedule of reading, make sure you have sick days and vacations. You're going to get sick (and sick of reading). If you schedule it up-front, you aren't stressing over it.

I was told I had only about 6 hours of effective reading a day. I stuck with that; the remaining 2-3 hours, I tracked down cites, I organized files, etc. I might have had more time, but I didn't push it.

How competitive is your cohort? Can you start a reading group as well?

And also, 48 hours to write a 50+ page paper? That's hard-core. I had 4 days to write 12 pages.
posted by printdevil at 12:20 PM on February 3, 2008

As it happens, I offer graduate courses in political science. American and methods, for whatever that's worth. Here's my standard advice.

(1) There is no better guide to this exam than previous exams. Most departments will let you see them in an official, above-the-board way. Ones that don't are idiots, because you can always ask people who've taken them in the past few years what were on theirs.

(1.5) Practice writing answers. Figure out the sorts of notes that are useful to you by noticing the sorts of things that you find yourself scrambling to locate.

(2) The second best guide are the relevant profs. Ask them what they think you should be focusing on. This is especially true of profs who might write questions for your exam but who you've never taken a class from.

(3) This advice was given to me, and is not remotely a joke. You should have some idea what you want to write your dissertation on, or at least what subfield you want to write in, and the profs probably know this too. Go and look at the APSR, AJPS, JOP, and top relevant subfield journal, IO or ISQ or whatever that is for you, and make sure you have read everything relevant in the last 5--10 years.

(4) Expect at least one vaguely methodological question on your IR exam.

(5) Have two or three relevant literatures ready to go to be used as examples in different questions. That is, one might ask you to talk about how some subfield demonstrates (or fails to) an accumulation of scientific knowledge, or one might ask you to describe some literature with a serious problem, discuss that problem, and offer a solution. You could easily use the same literature for both.

(6) Learning R well: use it. Work on your own project. Try to replicate someone else's results; any of the million articles that spin NES tapes are awesome for this even for non-Americanists. When you can't figure out how to make it do something, google that thing.
posted by ROU_Xenophobe at 12:26 PM on February 3, 2008

I learn which professors are writing questions for the exam six weeks ahead of time.

But you know who's in the IR and methods fields now.
posted by ROU_Xenophobe at 12:27 PM on February 3, 2008

Have you any reason to think that your outfit want to fail you? Most places don't -- they are out to keep all the good people they can, and even all the not-so-good-but-could-get-by ones. So look at the fail rate. Do you really think you naturally fall that low down the curve?

Don't go with a detailed plan -- just block out some targets, which you may well modify as you understand the field better. "By the end of the first month I will have read x of the key books in the area, and have a good idea about the main themes of another y", where x and y are suited to your speed of reading. (If you know it is relatively slow then schedule a speed-reading course asap, skimming is a very valuable skill.) Key texts probably don't need comprehensive notes -- your other reading will reinforce which were the important messages. The advice is always to make a database of the stuff you read, with a brief note about the contents. As far as I know, the advice is universally ignored.

Explore the background to your area. Don't treat it as something you must plough through from one end to the other, as people have said above you should concentrate on topics favoured by staff in your department. I am frankly shocked that you will be informed who wrote the questions, and the only reason that this info is relevant and being given to you is that it does make a difference -- they are not asking general questions, so don't waste your time on even coverage of everything. On the bright side, it is an indication they are giving you extra help to pass. It is very valuable to discuss stuff with other people -- profs and students, specialists and non-specialists. Don't let a schedule written 100 days before get in the way of seizing opportunities to digest what you have read by talking it through with others. Do spend time on your own particular iterests -- it doesn't look good if you don't have a thorough knowledge of what you want to take forward.

Learning to program takes time, effort and practice, and again you already have the advice above to use it for real stuff relevant to you. This learning cannot be done at the last minute, nor be completed early and then ignored. The rest of the test preparation does not need to be radically different from the way that has worked for you before.The last few days you obviously want to be writing practice answers so that they will flow easily when the real tests come. I see the 48 hours to complete the answers not as a terrible imposition but another way in which they are giving you every opportunity to succeed.

One useful thing you can do 230 days out is to make a commitment to eating a square meal every day, getting exercise every week, and taking adequate breaks, daily, weekly and every month or two. 230 days' hard slog would be a disaster.
posted by Idcoytco at 3:14 PM on February 3, 2008

Googling, I assume that you're at Davis. Definitely talk to Molly Melin at least about the IR exam if not both, since they should be relatively fresh in her head and she seems at least to be not evil. On the methods side definitely ask Molly or other people who are around what Jackman is looking for, since he seems like he can be awfully fussy when he feels like it. Given that Huckfeldt and Jones are newish, but can and probably will still write questions, you might google for recent Indiana and Arizona phds who are approachable (Patrick Brandt, Kris Kanthak, Bagel Johnson come to mind) and ask them what kinds of questions they tend to ask and what sort of answers they're looking for. I'm assuming you took game theory from Zeev... you should ask him whether you should be preparing to solve arbitrary games in extensive form, to prove classic results, or to discuss the problems of applying formal models to particular situations.

The big thing here, to answer Idcoytco's sentiments, is that you answer the question that they're asking in their own heads and not answering a different question, from a different base and with different presumptions about what good answers are.

Welcome to a small world.
posted by ROU_Xenophobe at 4:26 PM on February 3, 2008