Monday, December 12, 2005

Why the qualification exam doesn't prepare you for research

I'll admit, like most grad students, I bitch about my advisor (both of them) behind his back. "He talks too much." "He doesn't listen." Despite my complaints, he's a well-intentioned man who just suffers from the same social inadequacies we all do. He's a smart guy, and every couple of months he says something that makes a lightbulb go on in my head. Today, he lit up one of those 10000 Watt bulbs that light up Wal-Mart parking lots.

I won't go into the details, mostly to protect the identity of said advisor, but I will impart the lesson here.

THE QUALIFICATION EXAM DOESN'T PREPARE YOU TO DO RESEARCH.

For my department, the qualification exam is what students take to prove to their qualification committee that they have what it takes to do research. Years later, they take the prelim, which is a kind of oral contract between you and your PhD committee on what your dissertation topic is about. One year later, they defend and graduate. Of course, this all depends on the advisor, but it's a rough time-line.

In some areas of research, students are given two papers for the qualification exam. They have to read and understand the papers and their related works within a month, and give a stellar performance...er presentation to a group of three professors, none of whom is their advisor. The qualification exam is mostly about luck and attitude. It's about luck because you could get a committee of baby chicks or a committee of angry, blood-sucking pirates. It's about luck because you could get a good paper related to your research topic or a bad paper that consists of 61 pages of formal methods from hell. Ahem. It's about attitude because these professors can smell fear. If you cannot emulate a cocky, knowledgeable mini-professor, if you exhibit the least bit of doubt in what you are presenting, then you should just go home. I've seen very, very, very knowledgable people fail the qual and get kicked out of this department because they are too, well, mousy. It's stupid, but maybe being a cocky bastard (or at least being able to emulate one) is part of academia. Or maybe it's another facet of this bizarre culture.

In other areas, students are given a written test which they pass or fail. I won't speak on this because I don't know enough about it. I just know that if I had a six hour test to take on theoretical computer science, I'd likely wet myself.

That said, I came into graduate school having worked in Industry for four years. In Industry, it takes years to make anything, mostly because your days are filled with regression tests, writing white papers, attending meetings about 401K plans, and attending workshops about how to use Clearcase. Most of the time, what you made isn't particularly new, but what you made is QUALITY because other people have to use it. For some companies, this is truer than for others.

Three and half years ago, I started graduate school. I did the coursework. I got the grades the department demanded. I courted and begged for an advisor. I took the qual.

How did I pass? Well, technically, I didn't really pass. I got a conditional pass because some blood-thirsty pirate starting asking me R-A-N-D-O-M questions in the last twenty minutes. I had to take a course, but I passed. So, how did I get the conditional pass?

1. I read and read and read and read and read. I read many of the citations. I read book chapters that the author had written. I read other papers that criticised the papers. I read and read and re-read the papers.

2. I practiced and practiced and practiced. Including the first practice where I broke down and started crying in front of all of my friends. My girlfriend had to lure me back out of the bathroom saying, "Just finish it this time and it will get easier." Including the first practice where the beautiful man who would later become my boyfriend told this particularly enthusiastic though overbearing guy to shut up because, "She needs to answer the questions."

What did I learn about research? What did I gain from my experience in Industry and the qualification exam that would help me do research?

Nothing.

Research isn't about reading. Certainly you need to be familiar with related work or you will suggest the Mousie-Kitten model for web interfaces and people will laugh at you. Research isn't about making release 1.0021, making sure to refactor carefully and be backward compatible with old ideas. Certainly it helps if what you are doing works on Linux, works with the http, or works with AADL, or works with ______ so users don't have to have take for granted your weird-o setup to understand your results.

NO!

Research is about what I think. Lately I've been thinking to myself, "I'm just making this up!" But I've gotten to the point in life where what I "make up" isn't bullshit I'm writing on short essay test questions anymore. No, what I "make up" are my ideas and I have the intellectual back-up of ten years of working on this stuff to give these ideas weight. Yeah, I have to make my ideas work. Yeah, I have to show a room full of experts that my results are reasonable. I have to have my own ideas. I have to be brave enough to stand in front of people and say, "This is my idea and it is good so nah nee nah nah nah."

I don't know how to fix this. I don't know how to change the system so that it doesn't take other graduate students A YEAR AND A HALF AFTER THE QUAL to understand the concept of research. I'll think about that and let you know.


PS. You've caught me at a moment of anger combined with confidence. The usual angry and depressed FCSGS should be back in future posts.

3 comments:

DC said...

You know, I've only recently come to the same sorts of realizations. I think a lot of the problem comes from the other grad students in your group not helping you to figure it out. We've recently had to slap down the timid and not-well-thought-out ideas of some 2nd years in our group and it really pointed out to me that I knew EVEN LESS THAN THEM at that point. And I shudder to think what another year and a half might do to my perspective.

Unknown said...

My qual was one of those 6 hour dealies and so was Mr. Geeky's. 6 hours of theoretical computer science. For me, 6 hours of Renaissance Literature. Note that my dissertation is not on RL, but we had to pick a lit time period.

That having your own ideas thing--kind of scary. And you're right, they should get to that during or before the quals.

Thomas Paine said...

Interesting post. You should read the book "Disciplined Minds". It has an interesting take on quals, and what it takes to pass them.

Whether you agree with it or not, it'll make for an interesting read.