- Handbook for Graduate Students Preparing to Deposit
- Information on Copyright Law and the Graduate Thesis
- Qualifying exams
The Committee Chair is always from the student’s department. Usually, the Chair is also the student’s Adviser and gets promoted to Chair when the prelim (or proposal defense) committee is formed. Most of the time, the Chair is also the Director. Exceptions occur for example when the Chair goes away on sabbatical or the research takes a direction in another department. The Graduate College officially convenes the committee, based on the Chair and Department recommendation, but usually the student has talked with all the committee members ahead of time.
A good research question is actually the answer to a host of other questions, such as “Who should be on the committee?,” “What research method should be used?,” “How many cases are needed?.” or “What role should the researcher adopt?”
Just as a practical matter of getting done quickly, I’d avoid things that come across as criticizing the field or simply relating a personal account. This has nothing to do with importance or even method, but framing things in a way the committee, and people in the field, will understand easily.
Proposal and final examination procedures
Both the proposal defense and the final defense of the dissertation are two-hour exams. In the beginning, the committee usually meets for a few minutes while the student waits outside, just to see if there are any big issues that need to be surfaced early. Then, the student presents for around 20 minutes, using handouts, web resources, or slides as appropriate. Then each committee member gets to ask questions, after that a general discussion, then the committee meets again for a few minutes while the student waits outside. Finally, the student returns to hear a report on the results and any special things to change or do in the research and writing.
There comes a time to settle on one outline and work within it without expecting that it will be perfect. Dissertations often end up with an explicit structure that doesn’t quite match the implicit structure of the ideas. That’s OK, because it reflects the learning and discoveries made along the way. A revised structure can come in later, in a journal article, or book.
Below is the standard, familiar, ordinary, and probably easiest form to follow. Essentially, it is a way of setting out a problem (parts 1 and 2), a way of solving that problem (part 3), the solution (part 4), and reflections on that solution (parts 5 and 6). A structure similar to this is used in many dissertation and research paper competitions as a guide for evaluators. No one says that the dissertation has to have this structure, but it obviously helps the evaluators when they can easily find sections corresponding to the ones they need to rate.
- Introduction — tells what this is about; why it’s important; states research questions; summarizes the study
- Background (aka conceptual framework, literature review; theory; previous work) — sets the theoretical and research context; what related work has been done; why haven’t these questions already been answered?
- Methods — restates research question(s); what you did (as researcher); why this site; why these data? why 3 cases? limitations? etc.
- Results (aka Findings; stories from the setting) — What you found out; what happened. Sometimes this chapter is broken down into several sections or even separate chapters. Some people separate the descriptive parts from the more interpretive parts. For example, you might have three parts corresponding to the three case studies.
- Discussion (aka interpretation) — what does this mean? How does it connect with chapter 2?
- Conclusion — Briefly summarizes, but more importantly, synthesizes the results in terms of what this means and why it’s important; does this at a higher level than chapter 5; usually includes sections on implications for teachers, for researchers; open questions; sometimes 5 and 6 are combined into one chapter
- Bibliography — Should the bibliography for the dissertation include only the references cited in the text, or should it reflect the range of reading you did to get there? Generally it includes only those references actually cited. If some readings were significant for you then they would most likely work their way into a discussion of the theoretical and empirical background, and then be cited and included in the list. You could append “additional readings”, “related works”, or the like, but I don’t know that it adds much for a dissertation in the way it does for a book.
The proposal is usually in the 50-75 page range, but there’s no set length; some are shorter and some much longer. It usually covers the introduction, background/lit review/theory, and methodology chapters, plus the bibliography, but there’s no requirement that it be organized exactly in that way.
Dissertations and books
Should a dissertation be written like a book? A well-written book is fun to read, it’s ready for publication, and it demonstrates a command of the research area. But remember the rhetorical purpose for a dissertation, namely, to convince the committee that you’ve defined a research question, you’ve read and synthesized the literature, you understand methodological issues, you collected and analyzed data or textual sources well, and so on. When asked to judge a dissertation, committee members, despite their professed and sincere interest in reading good research prose, sometimes appear quite conservative in terms of their expectations regarding design. Also, in order to meet the kinds of questions committee members (or journal editors, potential employers, etc) ask it’s sometimes just a lot *easier* on you to rely on conventional formats, or at least, one that permits a translation into a conventional structures. The rule of finding a design that suits the content is probably the best one.
The standard dissertation format manifests one version of the inquiry cycle:
Viewed this way, we can see that the Introduction poses a question or problem to be resolved; Background lays out the results of previous investigations and a framework for this one; Methods describes what action is undertaken, what new is created; Results is the way that this study communicates to the research community its findings; and Discussion/Conclusion is a reflection on the research.
Problem/solution as hour glass
Another way to think of this is as an hour glass. If you conceive a dissertation as having the standard parts shown above, then you can see a broad/narrow/broad movement in this way:
The Introduction starts with what matters most in the grandest way. That’s where you’re allowed say things like, “What could be more important than how people work together to create knowledge? The web is the way we do that now, and thus, I’m interested in everything about the WWW.” But quickly, you need to narrow that broad vision down to a topic, then to an issue, and finally to a question.
In the Literature Review (or Theory, Conceptual Framework, or Background section) you look at how others have conceived this question, what they’ve done towards answering it, and what’s missing. So, out of all the things you could say, you’re continuing to narrow down. In the Methodology section, you focus even more–on a specific method for research, with specific populations, technologies, or information systems.
Coming back out, the Results mirrors Methodology in telling the readers exactly what you learned out of your specific research. In Discussion (or Interpretation) you expand further as you ask what it all means. finally, in the Conclusion, which mirrors the Introduction, you pull back to the broad view as you discuss connections, extensions, implications, and future research.
At the neck of the hour glass is your actual study. In relation to all the text around it, it may seem tiny. Your case study of six users, or your simple interactive system barely begins to address the grand issues of the Introduction/Conclusion. But it’s manageable, and has reasonable prospects of leading to concrete results.
The task of writing the dissertation then becomes to build a coherent whole, which situates the particular study or analysis within the larger frame. Moreover, it needs to show why and how this one investigation does in fact contribute to knowledge in the area, so that the Results/Discussion/Conclusion is not only a mirror image of Introduction/Literature/Methodology, but also is somehow extended in a useful way. You could think of the Results/Discussion/Conclusion as a move from data to information to knowledge.
It also needs to hang together. That’s why it’s almost impossible to say, “I’m all done with the literature review and ready to move on to another chapter.” As Donne says, “each section is a piece of the continent, a part of the main.”
Dissertations that are scholarly investigations may not have this structure, but they have a similar broad/narrow/broad movement, one that essentially thinks global, but acts local. The key here is not which sections (I’d quickly toss the structure I suggest above). Instead, it’s more about whether this is all pointing to a coherent whole, which articulates both some important problem, and some nugget of investigation within that.
After writing these guidelines, I realized that my own dissertation, The Logical Structure Underlying Temporal References in Natural Language didn’t follow them. Or, if it did, it illustrates how one needs to apply them appropriately.
Because mine had a structure that was part linguistics, part logic, and part artificial intelligence, it didn’t fit the APA-style outline, although each part was essentially a self-contained version of the problem/solution format. This approach is more common in the sciences and in Scandanavian countries. In some departments, a student may publish a set of articles and then compile them with an introduction and reflections into a single dissertation.