Roman's Not-A-Blog

Sporadic ramblings and interesting information.

Checking the narrative of a thesis/dissertation

In my life, I have assisted several friends in the preparation of their doctoral thesis / dissertations. In doing so, I have found that there is one particular element that is usually underdeveloped: what I like to call the "narrative" of the thesis.

I like to think of research articles and doctoral theses as narrative pieces. As such, they must follow various storytelling rules. Provide structure. Keep it simple. Foreshadow. Do not leave the reader hanging. Avoid inconsistencies. Do not be boring. In fact, I firmly believe that a thesis must kindly guide their readers through its contents. Yet there are always inconsistencies, impossible jumps from Section A to Section B, plots that are left partially unsolved. This is normal even in the final draft versions: writing a thesis is a daunting task, and it is difficult to revise the 400 pound gorilla once it is written (and revised, and revised again...). Well, writers need editors, and Ph.D. students need their editors too - to check those small details.

Therefore, this post provides various techniques that will help Ph.D. students to become their own editors. Tools that will help tired students to tame the 400 pound gorilla before it runs wild. Or using simple words: tools that will help to detect inconsistencies, tools that point out places where the narrative could be improved.

As usual, my disclaimer: The use of these techniques is entirely at your own risk. I will not be held responsible for any use of the contents of this blog post. Another extra disclaimer: probably these techniques are either i) already invented or ii) plainly obvious :-). Yet these techniques have helped some people. I will put them here in case they can help even more people.


WHAT: This technique helps to detect inconsistencies at an structural and narrative level, plus points out content that has not been properly introduced or is underdeveloped.


  • Summarize your thesis in a powerpoint presentation. This is not a draft version of the presentation that will be used during the defense: this is a quick hack draft that must follow the actual structure and contents of the thesis. The order must be respected (the slides must follow the structure of your thesis), nothing must be added (if it is not in the thesis, do not add it), and the presentation style does not matter (do not lose any second on designing the presentation). Then, defend it against an imaginary committee.

WHY: If a thesis have some structural or narrative problems, the researcher will find them, consciously or unconciously. There will be "a nagging feeling" when preparing or presenting the slides. During the (fake) defense, there will be problems when going from one slide to the other. Previously unexplained concepts will jump out of the slides like wild Pokemons.


WHAT: This technique allows the researcher to analyze how the aims and objectives of the thesis are fulfilled.

HOW: Fill up this table, one row per aim / objective.

Aim / Objective Rationale Result
Describe the aim / objective, including the page number where it is introduced. Succintly describe "why" the aim / objective is in the thesis. - Add an item every time the aim / objective is partially / completely fulfilled. Include the page number in every item.

WHY: By checking this table, it is possible to detect whether a group of aims / objectives are being properly presented and fulfilled or not. By including the page number in every item, it is also possible to check whether the results are presented in a logical order.


WHAT: This technique helps to detect potential issues and inconsistencies that might hinder the comprehension of the thesis.


  • For each paragraph, the researcher must ask himself/herself the following question: "WHAT is the problem described in this paragraph?". This helps to identify doubts and issues that the reader expects to see resolved at some point. These items must be written down in a document, including the page number where they first appeared.
    In parallel, the researcher must also ask himself/herself the following question: HOW are the problems in my list solved here?". That is, if the paragraph provides a partial or complete answer to any item listed in the document. If so, the item needs to be updated with the page number, alongside with a very short and clear description of the answer, including the level of fulfillment (Partial / Complete).

WHY: After reviewing the document, any item that is not solved or that takes too long to resolve (e.g. a basic background problem that is not explained in the introduction, but in a late chapter) is a symptom of a potential inconsistency.


WHAT: This technique is related to the "What" / "How" technique, but pursues a different objective: to discover if there are concepts or contexts that have not been supported by research or related references.


  • For each paragraph, the researcher should check if the text contains statements such as "We chose this X (instead of Y)", "A is better than B", "We propose this architecture", etc. And for each of them, the researcher should ask himself / herself the following question: "WHY?". The answer to this question should be near the paragraph that is being analyzed, or in an easily accessible place.

WHY: As a scientific text, any claim must be properly supported: by references, by analyses, by experiments, etc. Any unsupported statement is usually the symptom of a major problem. Note that applying this technique takes time, but we recommend to use it if there is time available. Any unanswered "why" is a weak point that might sink the thesis. Any "why" that takes too long to resolve hinders the comprehension of the thesis. Note also that there must be a very compelling reason behind all uncorroborated statements.