a journal of a researcher

Monday, October 11, 2004

How to Guide Master Students into Research

My experience so far told me that the master students are difficult to study a problem from abstract level at beginning. If you throw a problem to them, they just get lost. You need to guide them into the domain smoothly. Here is my recipe:
  1. Coding. Good researcher should be a good programmer. If your student can’t code, then, better do not work with them. Present the students explicit user case. Ask them to code a prototype system in a very simple way, without the research issues. This prototype is going to be the test bed of the research result later.
  2. Background investigation. Let the students know that width-breath background research is one aspect of research. It helps them to know the state-of-the-art. But this investigate should not be trivial. They have to read literatures. More important, hands-on on existing methods, i.e. try out the software, samples, coding.
  3. Select a problem to solve. This problem should be explicit, isolated enough. If it isn’t, you have to help the students to make it explicit and isolated. Then the student or you present the solution. The student tries out the solution, modifies it, improves it, and presents the results to you.
  4. Requirements on the proposal. The presented solutions should be tried and proved feasible. A proposal is not the beginning of research, it is at least close to 1/2 of the research. Thus a proposal is something you can really follow.
  5. Well motivate the students. Keep the enthusiastic in the students. Lead them into research without pain. Well guide the students, so that they can see their progress.
    Well control the students. You need to give students pressure. You need to point out the goal they have to reach before they can get their degree. Do not be too polite. Give orders explicitly and firmly. Keep your authority on the students.
  6. Demonstrate high quality of work. You need to show them that each step on research needs to be done in the first class quality. Do no tolerate compromises. Ask them redo their work, if they do not reach the quality.
  7. Work in group. I found they work much better if they are assigned the same work. Competition plus help each other.
  8. Thesis Writing. Ask them to accumulate materials for their thesis from beginning. Even draft some chapter while they work on the problem.


Some common weakness:

  • in the proposal, the problem is not explicit. The student still does not know what is the problem he/she to study.
  • Weak coding ability. The student has little real world experience to deal with complex problems. If they does only the assignments, you can be sure that they have not enough experience for coding. And you can not expect they can do research alone.
  • Many students choose to do thesis, instead of the other options, 3 courses and a report. You have to tell them that the actual work load of the thesis should be more than 3 courses and a report. Make sure they have good plan, do not let them sleep or rest.
    So this gives you an idea how hard you have to work. Do not believe you can sleep and rest.

0 Comments:

Post a Comment

<< Home