HCI Reading Seminar
CMSC 838B
Fall 2003
Prof. Ben Bederson

How to read a research paper

When you read a research paper, your goal should be to try and understand the contributions and limitations of the paper. Read it critically, remembering that a person wrote this that has a particular reason for writing the paper. They may be trying to push a particular approach, or technology. Even if they are trying to be as even as possible, they are likely to introduce biases. So, I have developed a short list of questions I try to keep in mind whenever I read a paper. We'll try to ask these questions when we discuss the papers in class.

• What is the problem (specifically what tasks does it solve)?
• What assumptions are made?
• Who are the intended users of the research?
• Have those users been involved in the design or evaluation of the work (i.e., is the solution usable?)
• Is the solution scalable (how much data does it work with)?
• Is the solution generalizable (does the solution work in other domains)?
• What is your favorite one line/sentence in the paper?
• What is wrong with the paper?
• What does the reference list show about the biases of the authors?
• What would have been a better and more descriptive title for the paper?

For this seminar, when you lead a paper, you should also read a few of the primary related references. We'll have about 25 minutes to discuss each paper, so you should plan on making a 7-10 minute presentation about the paper. I will cut you off at 10 minutes. In your 10 minutes, you should give the paper some context by briefly presenting the related work, and explaining how this work relates to previous efforts. And, you should briefly attempt to answer the above questions. You should not give a summary of the paper. You should assume that everyone has read the paper, and summarizing the paper wastes everyone's time. Rather, bring out the big issues, and open questions.