Marian's Ten Top Tips for Research Students

Compiled by Dr. Marian Petre, March 1998

Prepared for the 1998 SIGCSE Doctoral Consortium


  1. Read, read, read. Seasoned researchers typically have an evolving 'reference set' of around 100-150 papers that forms the core of the relevant literature in their specialty, and with which they are conversant. Students need to read enough to form an initial reference set.
     
  2. Write, write, write.
  3. Keep an annotated bibliography. This is the single most powerful research tool you can give yourself. It should be a personal tool, including both all the usual bibliographic information, and also the date when you read the paper and notes on what you found interesting / seminal / infuriating / etc. about it.
     
  4. Form an 'informal committee'. Try to find a small set of reliable, interested people who are willing to read for you, comment on ideas, bring literature to your attention, introduce you to other researchers, and so on. They may be specialists who can provide expertise on which you can draw, or generalists who ask tough questions.
     
  5. Expose your work. Make your work public in technical reports, research seminars, conference papers. The best way to get information is to share information; if people understand what your ideas are, they can respond to them. Making your work public exposes you to questions and criticism early (when it can do you some good), helps you to 'network' and gather leads, and gives you practice articulating your reasoning.
     
  6. So what? Learn to ask the other questions. Students often get a result and forget to take the next step. "Look, I got a correlation!" "So what?" Learn to go beyond your initial question, learn to invert the question in order to expose other perspectives, and learn to look for alternative explanations.
     
  7. Never hide from your supervisor. 'Hiding' is a pathological behaviour in which most research students indulge at some point. Communicating with your supervisor is pre-requisite to getting the most out of your supervisor.
     
  8. Always make back-ups (and keep a set off-site). More than one student has had to start writing from scratch or to repeat empirical work because he or she neglected this most basic of disciplines.
     
  9. Read at least one completed dissertation cover-to-cover. Reading something that has 'passed' is an excellent way to reflect on dissertation structure, content, and style - and on 'what it takes'.
     
  10. A doctorate is pass / fail. Part of the process is learning when 'enough is enough'.


Page maintained by Vicki L. Almstrum. Department of Computer Sciences at UT Austin
Send suggestions, comments to almstrum@cs.utexas.edu