Month: September 2015

Image based on Paltridge’s “Four Thesis Types”, cited by  Niall McMahon.

In this series of posts I outline a variety of different ways to structure your PhD thesis. Decide on a structure for your thesis early on because it helps you to plan your reading, technical and practical research and your writing.  Scroll down for more blog posts in this series that go into more detail about how to develop good writing habits and how to work with your supervisor.

The thesis IS the PhD – it is all the examiners have to go on. They do not know you. They don’t know, or care, how passionate you are or how hard you have worked. Every PhD thesis emerges through hard work and passion. In summary, to the outsider – your examiners –  all that really matters is the thesis.

To create the diagram shown above, McMahon analysed eight dissertations, taking into account Paltridge’s analysis of “fifteen doctoral theses from a range of disciplines and the advice given in eight recent publications from around the English-speaking world”. The link has many useful references. I have never supervised or examined a thesis that uses the Compilation Based structure in the arts and humanities. However, it as it is fairly common in the sciences, including computer science.

Regardless of the particular thesis structure that you choose, it is good practice for PhD researchers to produce a draft structure in the first few weeks as it helps to focus your thinking and can be used to prioritise your reading and writing.  It’s not a contract, but a ‘live’ document that alters over time. Having a structure mapped out helps both supervisor and researcher make sure the research is on track. Take a one page bullet point draft of your proposed thesis to your supervisor to discuss your plan of action. If you plan to add or drop a key theory or topic you can go back to your structure and see how that might impact your argument. You can also use this one page document and/or mind map of your thesis to find out of there is something missing that your supervisor is expecting. Use it to get feedback from your supervisor about whether or not a particular avenue is worth your precious time. But remember, it’s your PhD. You don’t have to agree with them!

Homework: If you have not done so already then prepare a structure using bullet point lists or a mind map. Schedule a meeting to discuss it with your supervisor, and send them the one page plan a week in advance of that meeting. 

See also “Hints and Tips on (Science and Engineering) Bachelor’s and Master’s Thesis Writing” by Peter McMahon, May 2009. I came across this while searching for Niall McMahon (note they are two different scholars who share the same family name). Peter McMahon’s full text here. Although Peter McMahon is writing about Master’s thesis writing his tips are very useful to PhD researchers.

This is the structure most often used by those who have built their research around case studies or iterative cycles of Action Research. How do you decide how many case studies to conduct? I think it’s a little like flower arranging, odd numbers seem to work best. The total number will depend on how complex and time-consuming each study is and the methods that you use. To get into enough detail and data analysis of a complex field it can hard to complete five studies, but on the other hand, doing only two prevents you from triangulating and from developing and showing your use of reflexive research methods fully. Three seems like a minimum. Some researchers do seven studies and those of you working in Human Computer Interface may complete many for cycles of testing that are sometimes grouped within a larger study.

In this structure, research methods are introduced and detailed via a standalone chapter, The General Methods chapter, and then discussed within each of the three studies. The General Methods chapter, before the chapters on the studies, is especially important if you are using a mixed method. You must refer to the method again in each study chapter, ‘bringing it to life’ by showing it in action in your research.  Use examples, with the appropriate language and terms from the method(s), to show the examiner, in detail, how you have deployed the method in each study and what, if anything, changed between studies and your application of the method.

It’s best practice for all PhD researchers to publish in peer-reviewed journals throughout their research period. There are many reasons why, as I explain in another post. In an earlier post I pointed out that you should make it a priority as soon as you start your PhD to identify key journals and conferences in your field.

The compilation thesis, also referred to as the multiple-manuscript thesis or anthology, is most commonly submitted by researchers in science and engineering disciplines. To produce a compilation-based PhD thesis you compile your series of published/publishable papers  into a coherent whole, with introductory and concluding chapters.  The introduction to the methodology employed is a substantial chapter typically about 50 pages long. It puts the subsequent papers into context and, in the case of multiple authors, established the candidate’s contribution to the published work. Note that if not the sole-author the candidate should be the principle, or first, author.

Different universities have different regulations about whether all these papers must have been published. Check your specific university regulations. Note that ‘publishable’ means the papers do not have to have been published but must be of a standard to be published in a refereed journal. However, there are many advantages to getting them published, as previously argued. If you are arguing that unpublished papers are publishable make sure your supervisor / committee agrees that the papers are of high enough quality. None of them need be published. All that is required is that they be publishable.

They should be sole-authored. Be wary of being one of many authors if you want to use the compilation-base approach to the PhD thesis. Jointly authored papers included in the three in your thesis may call the examiners to question who the author of the research is, who has made the contribution to knowledge.

Planning your work under the ‘three papers’ model is straightforward but depends on a good standard of writing. I have worked with computer science researchers who do not develop the habit of writing and redrafting that is necessary for writing peer-reviewed papers. Then they scramble to get three papers published. Develop the writing habit, polish your writing skills and writing these papers will become less stressful.

If you are in a Hong Kong or UK PhD programme then  you have three years to complete in full-time study mode. This tight timeframe for publishing is another reason why you must identify the key conferences and journals in your field and familiarise yourself with the writing style for those journal by reading and analysing papers that have made it through their peer-review.

If you being your PhD in September then you should aim to have your first paper finished by December of your second year, your second paper finished by December of your third year, and your third paper finished by September of your third year. In other words, you have 15 months to write your first paper, 12 months for your second paper and nine months for your third paper. This timing allows for the fact that you produce results more quickly as you go on, because you do not have to spend so much time cleaning your data, doing background reading, and so on.

For useful tips on this type of thesis see this see this PDF by Sarantis Kalyvitis, Professor of Macroeconomics and International Finance, in the Department of International European and Economic Studies, AUEB