Friday, January 21, 2011

Teach Me How to Research

How many of you were formally taught how to research? I'm guessing that for many of you, like me, there was never a class or workshop available to show you how it's done.  I wish there was.

Who Knows What You'll Find When You Ask Questions
Who Knows What You'll Find When You Ask Questions by [F]oxymoron

If I had ever known I even might go to grad school, I would have tried to learn more about research before I even got here.  I could have done a summer research internship with a professor in our faculty.  I could have tried to make my honours project a little more research oriented.

When I started my Masters, I had no idea what "real" research was.  Some of my courses tried to show us, but I usually ended up doing implementation-based projects instead of attempting to answer any interesting questions.  I learned a bit from my supervisors, but we never explicitly talked about how things are done, so even then I was kind of just guessing.

Something I think would help me - and I'm sure many others - is an optional class or workshop series on how research is done in our field.  What exactly constitutes a 'contribution,' and how do you know if you've got one? What do theory-oriented researchers do? How are experiments usually run? What are the different kinds of papers that we could write, and how would an algorithms theory paper differ from, say, an experimental paper in graphics? It would be great to get specific to our areas, too, because even if I'm not an algorithms theorist, I think it would be very useful to see the big picture by knowing how various computer scientists work.

Does your school have a course or workshops on how to research? How well does it work? If you have a good model I'd love to hear about it, and maybe I can suggest doing something similar at Carleton.

5 comments:

  1. There are mechanical approaches to research, i.e., definite steps one has to follow: Think of a topic, try to find out published literature about it, analyse and understand, then try to add some thing new to it. Thus, a new knowledge has been added in the existing set of knowledge. Hence, you have contributed to research.

    The other thing which has been counted as research is - giving a different interpretation to the work already carried out.

    In the first case when you add some thing new, first you give a hypothesis. Then, there is formal approach to prove the hypothesis or there is experimental based support, based on the data and conclusions from that.

    In very few cases the research is radically new, having no concituity with the existing work. For example, the Keplar's laws of planatory motions and Newtons law of gravitation.

    ReplyDelete
  2. The idea of a course on "how to do research" is intriguing. It's like having a chalk-and-blackboard or pen-and-paper lecture on "learning how to ride a bike".

    I believe the best way to learn how to do research is analyzing how others do research and trying it on your own.

    I understand that in the very beginning it is hard to know how to do research. In this respect, the most useful thing is taking classes that are "research oriented".

    In our department, we have tons of courses that involve reading at least 2-3 papers a week, write critiques on these papers, present and discuss one such paper during the semester, and participate in a (team) research project. The critiques and in-class discussions are particularly useful. You get to think about the published papers, their contributions, the problems they address, the research and experimental methodology, the related work, the follow-up work, etc. On top of this, the in-class presentations and discussions help you understand what others thought about the same piece of work that you read and analyzed.

    On top of research courses, we have reading seminars on different topics: computer architecture and compilers, security, networking, systems software, etc. The goal of these reading groups is either to read papers from top-tier conferences, present and discuss them, or present your own research, even if it is in its rough/early stage. Both these venues are extremely important in developing yourself as a researcher.

    As you can probably tell, I am all for a "hands-on" approach to learning how to research. Supervisors are very important in this respect as well.

    Good luck with your work!

    ReplyDelete
  3. I always like hands-on approaches, and what you describe sounds great!

    ReplyDelete
  4. Quite often the university library will have a subject liaison librarian that specializes in research for a specific college. They offer research consultations and often classes as well. Usually, at least in my experience, they are underutilized resources. They would address at least part of the problem. Still they wouldn't necessarily address the "contribution" parts of your question. :)

    Jeremy

    ReplyDelete
  5. Quite often the university library faculty will offer both research classes and consultations. Usually, there is a designated librarian-liaison who specializes in that particular departments subject matter. They often create guides and finding aides related to conduction good research in that particular field. I've found that they're often an underutilized resource. They often also have additional degrees related to the departments that they are responsible for and can provide a bit of an inside perspective. I think they would be a good place to start, although some questions, such as what constitutes a true contribution to the field might need more expert (i.e. professors) input....

    Jeremy

    ReplyDelete

Comments are moderated - please be patient while I approve yours.

Note: Only a member of this blog may post a comment.