Papers in journals, etc., are published reports from ongoing research projects. The reason it can seem so hard to understand a single paper, to remember what it says, etc., is that each paper is not intended to stand alone. To understand a published report, you have to learn to "see past the paper" to the underlying research project.

All well constructed research follows the same general skeleton. While a single research project may not have all the elements listed below (for example, not all computing projects use formal research questions at this point in time, and exploratory research does not have formal hypothesis), a well constructed project should have most of them.

Note that the order is not significant! Also, while the vocabulary given here is very common, no global consensus exists as to usage concerning a few of these terms. What matters is the presence/absence of the elements, not what they are called. For example, you will often see the use of goals and objectives switched. In terms of clarity of planning and thought, what matters is that the researchers distinguish between the two, not what they call them.

Research Skeleton

• Problem statement: What is the problem you want to address?

• Research question(s): need to come up with a short description grounded in theory and subject to empirical investigation

• Assumptions (How are you going to bound your problem so that it is manageable? What decisions did you have to make for which you couldn't cite someone else's results or do your own research?)

• Goals (The specific goals you intend to achieve in this project.)

• Approach (How will you go about achieving those goals? Is there a particular theory you are following? Base your approach in what others have done or tried to do.)

• Significance (Why do we care if you achieve your goals? What impact will your completed research have?)

• Objectives (These are measurable objectives for each goal. They follow from the discussion of approach and significance.)

• Method (What specific methods will you follow to meet your objectives? Are you using experimentation? If so, what type? What data will you use or collect?)

• Hypotheses (What specific testable hypotheses are you going to validate in your research? The hypotheses should match up with the objectives and goals.)

• Data/Results (What happened? What statistical model are you using to analyze your results? What was the outcome? Did you validate your hypotheses?)

• Conclusions (Did you meet your objectives? Achieve your goals? Why or why not? What new knowledge arose from your work?)

• Future work (What will you do next? What do you think are promising avenues to pursue, given your results?)

What to do

When first reading research reports (of all sorts), a very useful thing to do is to create a blank version of the research skeleton for each report, and then fill it in as you go. Fill it in with quotes (don't paraphrase, because it is too easy to insert your ideas - things you think are there, but are not) drawn from the paper, as you go. Yes, it's great if you have people you can talk with about whether you are on track, but don't panic if you don't have anyone to check with to see if you are right. As you work through more papers, if you keep filling in skeletons, the process will self-regulate. We've done this a lot over the years with lots of students, and it works.

Ann Peterson Bishop's "Digital libraries and knowledge disaggregation: the use of journal article components", published in the Proceedings of the 3rd ACM conference on Digital Libraries (DL '98, pgs 29-39), presents the results of a research project concerning how student and faculty researchers disaggregate (decompile) what is presented in a research report. Reading her paper may help you develop your own process, as well as being an excellent example of a research project and a research paper.

[see also UnderstandingANewField]