Content Tips for your Dissertation or Project Write-up

These tips are mostly focused at writing an academic (and scientific) dissertation. But I think most of these suggestions are sensible enough they should feature in any lengthy document. Based on my experience as an examiner and a supervisor, these are the most common things I notice each time I pick a project write-up.

Here some tips to get you going:

1. Use a proper (official) template; this will go a long way in guiding you
2. Start with the ToC first, and show it to your supervisor.
3. Additionally, write a two/three-page document that describes the narrative of your work. This is for your own (and your supervisor’s) consumption, but will help you with the setting of the scene
4. Look at two or three “Distinction” grade dissertations, from the same department/course. This will show you what is acceptable in terms of the structure/contents
5. Use the active voice, it makes reading more interesting. Also, most journals require this as a stylistic choice.
6. Use first-person plural; i.e. we/our (e.g. We examine the performance of four freely available small molecule conformer generation tools.)
7. Most of the project should be written in the present tense (with the exception of the methodology mostly). Obviously, do not mix tenses in the same sentence!
8. The abstract should be a summary of your work (and not a summary of the background). It should include why this work is important, brief methodology, your results, and main conclusions. There are many tutorials on how to write a good abstract. The abstract should be written last.
9. Each chapter should have an ending “Summary” section (note do no name it “Conclusion” there is only one conclusion section), with the exception of the Introduction and Conclusion. The Introduction chapter should have a “Document Structure” section instead of the “Summary”
10. Consistency, above everything else
11. For the “Introduction” chapter, “Aims and Objectives” should be clear and explicit. For scientific dissertations, the aims may be formulated as research questions. Objectives should not be implementation details
12. Objectives should reflect the major experiments and findings in your work. This means that the rest of the dissertation should follow the narrative set in the objectives (there should be subsections in your methodology and results to match the objectives).
13. The objectives are not steps in your methodology
14. The methodology section always starts with a schematic of your approach and/or architecture, and continues with an explanation of each of the processes in this schematic/diagram
15. No contractions (e.g. won’t)
16. References must be consistent with each other and in the required style (e.g. year at end, full or abbreviated names). I personally prefer Ebejer et al. than a numbered reference [21]. The reason for this is that examiners will be familiar with the main papers in the area, and they can verify that you cited the most important work (and you cited it correctly) without having to flip through the “References” section
17. Remove clutter from References, e.g. DOI. All referenced texts should have the same set of fields
18. Latin abbreviations should be in italic
19. For an M.Sc. from the University of Malta, your project is a dissertation (not a thesis). For a Ph.D. there, your final write-up is a thesis. This is dependent on your institution, but typically these words are not interchangeable. You can also use the word study or research project to refer to your work
20. Equation, Figure, and Table should be in full and capitalized if referring to a specific instance (e.g. Figure 3). Same for Section, Chapter etc.
21. Titles should be short (less than 12 words), and punchy
22. Your Title should not be hyphenated anywhere (certainly not on your title/cover page)
23. 0-10 numbers should be written as words, otherwise numeric
24. Numbers should have thousands separator
25. Numbers in tables should be right-aligned
26. You should not start a sentence with a number. If you do, spell it out as a word. If the number is long (e.g. 3,152), reword the sentence so the number does not come at the start of it. Tangentially note that many programming languages also do not allow the use of a number (e.g. 2) at the start of a variable name
27. Use a consistent number of decimal places when reporting your results (typically three in a scientific setting)
28. References have to be consistent with each other (e.g. year always at the end). Make sure to include pages and volume information
29. No references to Wikipedia and other non-scholarly articles (blog posts)
30. Each figure (or table) lifted from somewhere else should have “reproduced from Del Piero et al. (2010)”. If you have changed the figure slightly use “adapted from Del Piero et al. (2010)”
31. Make sure to adhere to the word limit. For M.Sc. in AI (ICS5200) this is 32,000 (but double-check with the course by-laws, as this may change). Your chapters should be distributed accordingly with this in mind. For example, for a scientific dissertation – Introduction (4-5k), Background and Literature Overview (7-9k), Methods (5-6k), Results & Discussion (6-8k), Evaluation (3-4k), and Conclusion (2-3k) for a total of 27-35k.
32. Capitalization (e.g. for section titles) is important; there are rules which govern this
33. Section headings should not follow each other immediately but you should always have some preamble text explaining what is coming next and why it is important
34. Cross-link between sections of your dissertation. It is fine to repeat yourself a few times across chapters (but not within chapters). Keep in mind that examiners will not read this in one sitting, but probably a chapter at a time
35. You should not repeat the same word(s) or phrases in the same sentence
36. Each equation/figure/table must be referenced from the text (and ideally placed closed)
37. A caption should tell you the whole story. A good dissertation can be read-only using the figures/tables
38. Learn what should be placed in a footnote (e.g. URLs)
39. For URLs (in footnotes) and other online resources, you always need the “Last Accessed” date
40. The mark of a footnote should fall outside of adjacent punctuation (such as quotation marks, a comma, or a period). So, footnotes at the end of the sentence should be placed after the fullstop.
41. Also, there should be no space between the word (or punctuation) and the footnote mark following it
42. Take care your paragraphs are not too long. Each paragraph should be made of a few sentences and should convey (only) one point. Also, take care your sentences are not too long (and without many sub-clauses).
43. Scientific writing is terse by definition. Remove useless words like very and actually. More so, the word very is for lazy people. Avoid it. There is always a better word than very something, e.g. very tired = exhausted
44. Related to the above, there is this beautiful list of misused words in science. It is a shame I do not know the source of this.
45. If you use “First”, readers also expect “Second” and “Third” nearby
46. No hanging lists. Also related, no titles or headings at the bottom of the page, with no following text
47. Make sure to use proper grammar; which vs that; where vs were, etc.
48. Your literature overview should link back to your work, and not just be a report of unconnected work
49. Never leave the reader questioning, tell them your story
50. Avoid “above” and “below”, when referring to equations/figures/tables/sections/etc.
51. You should explicitly define and explain each term in an equation in the immediately following text
52. Keep regular backups of your writing, and in multiple locations. There is nothing more frustrating than finishing a chapter only to have to write it again because you lost it …
53. For your viva, take a printed copy of your dissertation
54. Always be specific when cross-referencing (i.e. to a specific section/chapter)
55. Make sure figures are of the correct resolution and easily readable (no too small, sharp, etc.)
56. Each time you compute a “mean” for a list of values, you also need a measure of spread (e.g. standard deviation, standard error, etc.)
57. With very few exceptions, there should be a space between a number and a unit (e.g. 32 GB). Ideally, the space is marked as non-breaking (e.g. in LaTeX 32\~GB)
58. The above point does not apply to percent signs, which should immediately follow a number (e.g. 50%)
59. When presenting numerical results in tables, make sure to highlight (e.g. bold) the best performing results, especially if comparing to other methods. This draws the attention of the reader to the important values
60. There should be no space between the last word of a question and the question mark (exact semantics of a period or fullstop)
61. One or two spaces after a period (or fullstop)? This is an open debate, which I think is made less relevant by modern word processing software. Still, …
62. Chapter/Section headings should never end with a period. (Thanks MBK!)
63. All chapters, sections, subsections, and subsubsections titles must be numbered
64. Page numbers; always (doh!)

Do you have more hints and tips for the budding dissertation writer?  If so drop me a comment, an email, or a tweet.