Thesis Writing Guide/Printable version
The current, editable version of this book is available in Wikibooks, the open-content textbooks collection, at
https://en.wikibooks.org/wiki/Thesis_Writing_Guide
Introduction
Abstract
- Context:
Writing a scientific bachelor or master thesis is complicated and only really understood after submission. Having these lessons learned available at the start of a thesis could lower the burden of writing it.
- Objective:
This guide is designed to wrap these lessons learned with best practices for every step of the thesis. It aims to be helpful to any reader looking to write a good bachelor/master thesis, from software setup over writing to final publication.
- Method:
This guide structures advice commonly given in everyday thesis supervision, alongside open subsections to compliment with individual experience at a very low threshold. First, the Introduction provides an overview over what this guide can and cannot do, preparing the reader to start with their thesis. The Setup represents the step 0, discussing aspects relevant before writing the first word or starting the first experiment. Work is in progress. Likely the most extensive section is Writing, detailing many different chapters and how to fill them with relevant, high-quality content. This written artefact is then submitted for Publication, a closing chapter handling everything after the final word was written.
- Result:
Since this guide is currently in the making, we don't have any noteworthy results beyond what the reader may find in the book. If the guide persists and is adopted to everyday academic supervision and writing, we see potential to replace this placeholder paragraph with a success story of how sharing knowledge helped everyone. To make this work, please use and share the book and contribute to the lesson learned sections. Contributing to the other sections or Discussion is also welcome.
- Conclusion:
This guide wraps lessons learned to write a bachelor or master thesis. It is a living document, used and extended in everyday academia. While not complete yet, it may help already, and invites for collaboration.
Specifics
How your specific thesis should and will look like depends on
- Region (e.g. Lower Saxony, Germany)
- Institution (e.g. Leibniz Universität Hannover)
- Faculty (e.g. Faculty of Electrical Engineering and Computer Science)
- (Sub)department (e.g. Data Science and Digital Libraries)
- Evaluator/Supervisor (e.g. your Professor)
- Time (e.g. 2024)
This guide attempts to unify all universal guidelines, while individual rules are separeted into modules:
Formal Organization
Then you need check a few boxes before your thesis can start:
- Get thesis admission
- Find the right supervisor (and evaluators)
- Settle on an assignment, preferably written out as a one-pager
- Find a thesis title, preferably agreed upon by you, your supervisor and every evaluator
- Register your thesis
This section will guide you through all of those steps.
Get thesis admission
Usually, you need to complete a certain set of requirements to sign up for a bachelor or master thesis. To ensure you have satisfied them, contact your examination office and get an admission sheet for your thesis (this may vary between institutions).
It is highly advised to complete the formal registration before step 2 and 3, to avoid conflict with the examination regulations.
Finding a topic & supervisor
Sadly, this guide has little to suggest for finding the right topic and supervisor. In an ideal world, there would be a marketplace like ebay, where any supervisor just lists their topics and any student just goes shopping until they find something, potentially even raising their own requests for topics. In reality, only few alternatives can be discussed in this subsection.
General advice
- You can search preemptively to get a feeling for what topics are available, even half a year before you actually start.
- When you contact a potential supervisor, make sure to
- make clear where you are in your studies and when you plan to start your thesis,
- provide a curriculum vitae (CV) and a transcript of results, alongside three short sentences describing what you're interested in,
- so they can provide you with a mutually beneficial answer.
Departments
While you study, you will listen to different subjects. If one is particularly interesting, just contact the lecturer and ask if they have topics available.
Online
If you're interested in a subject, you can just search for it on the internet, especially on dedicated science platforms like Google Scholar. If you search for the name of a few professors you know, you may find what they generally work on, and whom you could approach.
Other universities
Technically, even professors at other universities could be your evaluator, depending on your university rules. Make sure to get clearance at your own university before you approach the external professor.
Formalize an assignment
It is advised to formalize an assignment before you start your thesis. This assignment could look as follows:
Motivation Writing a scientific bachelor or master thesis is complicated and only really understood after submission. Having these lessons learned available at the start of a thesis could lower the burden of writing it. Goal The goal is to design a guide that makes this experience accessible to bachelor and master thesis students while writing their thesis. The guide needs to be openly available, expandable and provide general advice applicable to different topics and domains. Approach The guide will be written by crowdsourcing previous experience of several pre-selected student thesis supervisors, all of whom have written and supervised bachelor or master theses before. During the writing process, the guide will be available online and open for feedback and collaboration. To validate this guide, it will be offered to several students during their thesis. These students will use and asses the usability of the guide and provide feedback in a semi-structured interview. This feedback will be evaluated and summarized to formulate recommendations to further improve the guide.
This assignment has several benefits:
- It serves as an anchor point so you and your supervisor always know the goals of your thesis, which is especially helpful when some time has passed since you settled upon it.
- It is basically identical to what you need to write either way in your introduction chapter. You're not doing additional work here, you're front-loading future work.
- It serves as a ground truth between you and your supervisor. When your thesis starts, your supervisor will likely know more about your topic than you do, but eventually, that will likely flip as you invest months into a specific topic. This clear write-out gives you a stepping stone to easier understand what their keywords might mean, especially compared to just a short title. This greatly reduces the risk of a rude awakening when the title is "solving the moving sofa problem" turns out to be a 50+ year old complex problem.
However, some points require consideration:
- The approach should be formulated that it provides a clear, but not too binding guideline. For example, if the number of participants was predefined to, say, 20 students, and only 15 are found, it would put an unnecessary burden upon the thesis. Always read the assignment top to bottom again to spot those pitfalls early on. However, these qualifiers also provide structure, so providing too blurry of a definition (e.g. "The guide will be evaluated.") provides little benefit either. Find a good balance with your supervisor.
Formulate a title
The title of your thesis should
- be as short as possible and as long as necessary
- reflect the assignment in a few key words
- be something you want to list in your CV later
- not change throughout your entire thesis process
All of those are guidelines, and there are always exceptions. The most important thing is that you and your supervisors agree upon the title.
Register your thesis
Once you have your thesis
- admission,
- supervisor and evaluators,
- assignment,
- and title,
you can finally register it. This also sets the date of your final submission. Be sure to register as soon as you have settled on a title, again, to avoid conflict with the examination regulations.
Look ahead
Now you don't need to worry about formalia, your thesis lies ahead of you. Some points are raised now to avoid confusion later.
Submission
The submission date is defined once you register. Shifting it to the future is sometimes possible, e.g. if unforeseen events delayed progress through no fault of your own. Always contact your examination in that case.
Specifics of submission requirements vary by departments, be sure to check Specifics for details. If that section does not provide any applicable to you, ask your supervisor to add them there, or do it yourself after you've asked them.
Presentation
In many cases, a thesis is concluded by a presentation. This presentation, sometimes called defense, may be evaluated, sometimes even graded, again, depending on your department specific regulations. Later in this book, this presentation will receive a dedicated chapter, for now keep one thing in mind:
While writing, remember that you eventually need to present this thesis to a room of people who may know little about your topic. If you can, create graphics and visualizations that illustrate what you want to talk about as you write your thesis. You will be able to use them twice, in your written thesis and in your presentation.
This saves you from having the prefect image that could have shined on your first page of your approach section, but only during preparation for your presentation slides, long after you've already submitted the written thesis.
Lessons Learned: Introduction
Dear reader, this is your time to shine! Contribute your own lessons learned, add helpful hints and mistakes made so that others may learn from it. Just click "edit" and add your own subsection!
Talk with your supervisor
Ask anything
Go ahead and ask that silly question you have in mind - whatever it is, just ask!
Meet frequently
Even if you feel you haven’t worked enough or think your progress isn’t good, still meet with your supervisor, update them, and seek feedback.
Feedback is a multiplier
Regular, constructive feedback from supervisors and peers not only boosts motivation but also helps refine the methodology I am currently working with.
Done is better than perfect
Don't expect perfection
You may sometimes feel like you have to know a lot and be super prepared before you could start. Honestly, you don’t need to have it all figured out - just start! Don’t let perfectionism slow you down; making progress is what really matters.
Don't be too hard on yourself, but always try
The only way to know if you can do something or not is to try, get feedback and improve your work. Everything can be learned, as long as you put your mind to it. You may have to move beyond your comfort zone and learn new things, but you can take the opportunity to grow in a supported environment. If you've gotten to the point of writing a thesis, you can also clear this bar. If you're struggling, you always have the help of this guide and your supervisor to seek advice.
You're doing something new.
Whatever you do, keep in mind you're likely the first to do it. Sure, many similar things have been done, but if you've got a good research topic, nobody else has done your specific variation of it. So try to live up to the standards of your field, but keep in mind any scientific endeavor can fail. Often a well documented and described failed approach with much to learn from is better than a successful experiment nobody can learn from, reuse or reproduce.
Setup
This begins with a establishing a basic setup of Reference Management and Writing Environment. It concludes with practical Lessons Learned: Setup, contributed by individual readers who may or may not be co-authors of the overall guide.
Reference Mangament
TL;DR: On day 0, setup Zotero and bookmark everything you read.
Since references are often an afterthought, but are actually the most valuable currency in science, this point is raised first.
Some rules:
- Read ≠ saved.
You won't be using every source you read, but unless you can find them again, you can't even use them. There is nothing worse than knowing you read the essential point you wrote two pages about already, but can't find the right source anymore and have to rewrite or delete this work again. - You won't save too many.
Your thesis will likely have 100 or 200 references at most. Even if you save everything you come across, you will end with maybe 10.000 references in your library. Searching this filtered library for one specific source you read is always much, much easier then having to find it again on the entire web. The average user will save their 350 papers, documents and websites in a Zotero library, cite maybe 50 of them.
Zotero
ORKG
other
Writing Environment
TL;DR: Use Overleaf, preferably one hosted by your university.
Proceed as follows:
- Check if your supervisor has a template.
This allows you to don't mind some formal requirements that are already preconfigured in the template. - Set up a shared access working environment, preferably over Overleaf and shared acces via invite/link, or using a cloud service such as OneDrive or Google Drive.
This makes reviewing and providing feedback easier. - Set up a backup schedule.
Back up your thesis every few weeks to a second service, be that a local PC or a different cloud. You can recover two weeks of lost progress much easier than your entire thesis if your primary service is offline when you need it most, like in the final weeks before submission. If you ever feel like you can save time by not doing backups and this won't happen to you, just have a short look at history (e.g.[1])
If those 3 points are checked, e.g. if you have setup your Overleaf project using the university cloud service and LaTeX template, you can proceed to the next section. If you're looking for more details on particular choices, continue reading this one.
LaTeX

Either Overleaf (online) or Visual Studio Code (local):
- Overleaf
- Visual Studio Code + a TeX distribution
- MikTex + Strawberry Perl
See[2]:- Install Perl. You can use Strawberry Perl in Windows.
- If you don't have administrator privileges you can install the portable version and add the path to the executable to the PATH environment variable.
- Install MikTeX. The creator of LaTeX Workshop suggests to use TeX Live instead because it already comes with Perl and you could skip one step in this list. The disadvantages of using TeX Live instead of MikTeX are more (see here).
- TexLive + LaTeX Workshop
- MikTex + Strawberry Perl
Word
Working with Word in a scientific environment is not recommended over LaTeX, but sometimes the right choice in certain circumstances, particularly when your field or supervisor provide better support for word. This guide recommends giving LaTeX a try and does not further elaborate on Word. If you are a supervisor with Word experience, feel free to supplement this section.
Writing assistance
TL;DR: When in doubt, don't trust (AI) tools.
However, a few tools are recommended:
- Grammarly, especially as a browser plugin.
The freemium version suffices to provide some feedback for the broadest mistakes. It does not replace careful proofreading and consideration, and should not always be trusted, as it can make mistakes as well. - LanguageTool, especially as a browser plugin.
Works in tandem with Grammarly, does roughly the same.
Translation tools are also worth mentioning, for various use-cases. Primarily, most scientific writing is in English, which might not be your first language.
- DeepL, for translating.
Keep in mind that you should never translate large portions of text just for copy-paste, but rather create your own translation and check with DeepL where you are uncertain. DeepL is also especially helpful to find synonyms or rephrase a sentence. - Linguee, for translating individual words and see in which context they are used.
ChatGPT
A dedicated section would be worth to discuss ChatGPT, or any LLM for that matter. In short, this guide does not recommend using ChatGPT in general. However, it is being used, see the "delve" phenomena[3]. As such, we need to provide some ground rules:
- Raise this point to your supervisor and get some feedback. Many faculties have their own rules and regulations, some particularly even for ChatGPT or LLMs in general, so just ask.
- When you use ChatGPT in any extent going beyond just finding selected reformulations or synonyms, like e.g. with any tool mentioned above, disclaim that you used it.
- Be aware that you want to be the author of your work. Every bit ChatGPT writes for you is like a colleague writing for you, which would be fraudulent in every other exam. Handle it with care and use it sparingly, and as mentioned above, when in doubt, don't use it at all.
Lessons Learned: Setup
Dear reader, this is your time to shine! Contribute your own lessons learned, add helpful hints and mistakes made so that others may learn from it. Just click "edit" and add your own subsection!
Don't setup too late or too long
It's best to start on your setup even before you actually start your thesis. Having your Google Drive, Zotero and Overleaf prepared a week or even a month in advance can help to focus on your task at hand and not stress to much about how 3 new applications work alongside solving an actual scientific problem.
Single root
Make one place your central canvas. This could be:
- A Miro board / PowerPoint / Google Slides / ...
- A stack of paper
- A OneNote / Obsidian / ...
- A real whiteboard
Make sure everything you do is somehow referenced from there. If you have a Zotero Library, link it there. If you have a Google Drive, link it there. If you send a mail, take a note there. If you have a meeting, write a protocol there. Put your thesis task on top and look up every once in a while.
Organization is key
There is (almost) no such thing as overprepared. Sure, you shouldn't spend more than 50 % of your time taking notes and sorting them, but anything below that likely pays of in the long run. Twice a week and whenever you're overwhelmed or need a breather, sit back for 20 minutes and structure your notes, sort your references, clean your library, shift sentences and bullet points around.
Writing
Before you write
Authorship
Gender neutral language
Citation
- If available, cite conference or journal articles zitieren, not (arXiv) preprints
- If you cite a preprint, write it in your text that "according to the preprint ... by ..."
- If available, use DOIs, not URLs
- never use both - if you have a DOI, you can delete the URL from the BibTeX
In recent years, a concept of "predatory journals"[4] has become more important. As a bachelor/master student, you are not expected to be able to differentiate them from "real" journals, but if you have a bad feeling, feel free to research the journal and publisher. You can always ask your supervisor for help as well.
If you need space
Usually, you don't need to worry about space in a bachelor/master/doctor thesis. If, however, you submit your work as a paper, or are otherwise restricted to a certain set of pages, here are a few hints for further cleanup:
- if the note says "note = {publisher: ...}" - you can replace it with "publisher = ..."
- a simple search & replace with
note = {publisher:andpublisher = {should do the trick most of the time, unless there is something like "number: ..." in between.
- a simple search & replace with
- If there are more than 4 authors, replace the 4th and all following with "others"
author = {First, A and Second, B and Third, C and Fourth, D and Fifth, E and ...}
becomesauthor = {First, A and Second, B and Third, C and others}
While you write
Pictures
Never copy-paste a drawing that you could re-create. Always cite where you derived an image from.
SVG > PDF > PNG https://www.drawio.com/doc/faq/svg-export-text-problems
Captions
Imagine you're curating a museum. Some poeple will only read the small sign beneath every picture and move on. Write short, descriptive captions that tell the viewers everything they need to know. Some readers will go "window-shopping", i.e. only view images, tables and their captions. Make sure your thesis is at least somewhat comprehensive with only those tools. Details can go into the fulltext, but the images and captions by themselves should be understandable already.
In short: Give ~ 3 lines of short, pointient description of what is shown and what is noteworthy. Be descriptive, what is shown, and keep the "why" and "how" for the long-form text.
Also:
- Table and Listing captions above,
- Figure and anything else below
Readers read Tables and listings from top to bottom, just like text, so captions should be above. Figures and most other inserts work differently from text, so captions below.
Introduction
Mandatory chapter, almost entirely derived from your assessment (see Thesis Writing Guide/Introduction#Formalize an assignment).
Background
Anything that could be interesting for this thesis guide. Even if it might not end up being used, it should be here.
Related Work
Approach
Implementation
Evaluation
Results
Discussion
Future Work
Conclusion
Lessons Learned: Writing
Dear reader, this is your time to shine! Contribute your own lessons learned, add helpful hints and mistakes made so that others may learn from it. Just click "edit" and add your own subsection!
Don't Procrastinate too long
If you push writing too far to the end of your thesis, you won't remember anymore what you did in week one. Start writing your master thesis in the first days. Anything is much better than nothing!
Scientific word choice is complicated
Structuring your ideas and using scientific terms is hard. Especially if English isn't your first language, you may not find the right words, but even a native speaker will have to re-learn a lot of words. "Significantly better" has a completely different meaning in science, compared to everyday usage.
In general:
- avoid adjectives, unless your citing them
- write what works for you, and ask your supervisor to give you feedback
- especially if you're missing a specific word, just mark BIG, BOLD AND RED AS TODO and just ask your supervisor
You will be a much better writer after your thesis, and that's good - you're learning, that's what all of this is for.
Take a walk
Often times, you can fix a writes block just by going outside and walk around for 20 minutes. Go for a walk often, it'll pay of one way or the other. In fact, after you read this, fix what you're doing right and then take a walk, feel for yourself :)
Think in headlines
You know everything that's going on, readers don't. When you write, start big and easy to understand. Then go into the next level of detail, complete that. Then the next level.
If you start on the highest detail level, nobody can follow and you will loose the reader immedieatly. Give them a bigger headline, a bigger picture first, then fill that in with more detail.
Chapter 2: Background
Anything that could be interesting for this thesis guide. Even if it might not end up being used, it should be here.
Background
Lessons Learned: Introduction
Dear reader, this is your time to shine! Contribute your own lessons learned, add helpful hints and mistakes made so that others may learn from it. Just click "edit" and add your own subsection!
Do not stray far
To write a good thesis, you need to read a lot of articles going in the same direction. Above all - do not stray far from your topic! You will soon find out how deep the rabbit hole goes, and your topic is specifically chosen to be manageable. Keep looking at what you have to do, not what you could do - each work has a "Future Work" section for a reason, write all your ideas down there and don't spend the time while you haven't finished your actual tasks.
If you're done and still have time, you can still work on anything from your future work list :)
Patience
Investing time in understanding the tools pays off in the long run. At the same time, don't overinvest from the get-go, maybe you read and write 5 pages about a method or tool which you are actually never going to use, because your supervisor has a better idea that you could just ask them for. As always, iterative steps are the key:
- Read, write down and structure as much as you can.
- Review with your supervisor.
- Refine your approach concept.
- Happy? Time ran out? Proceed! - If neither - back to step 1 :)
Chapter 5: Implementation
Lessons Learned: Implementation
Dear reader, this is your time to shine! Contribute your own lessons learned, add helpful hints and mistakes made so that others may learn from it. Just click "edit" and add your own subsection!
Test, test, test
Iterative improvement is generally much more successful than designing everything behind closed walls and never test it before the final evaluation. Be sure to test as much and as often as you can, document the process and build your development around it. There are even some (research) methods designed around this concept, e.g.
- Quality Improvement Paradigm
- en:w:PDCA
Start with small-scale experiments before scaling up. Early prototypes often reveal limitations that are easier to address before expanding scope.
Chapter 6: Evaluation
Survey
Design
- NASA TLX Score https://en.wikipedia.org/wiki/NASA-TLX
- Scales of Measurement - Nominal, Ordinal, Interval, & Ratio Scale Data https://www.questionpro.com/blog/nominal-ordinal-interval-ratio/ https://www.youtube.com/watch?v=LuBD49SFpWs
- Likert Scale https://en.wikipedia.org/wiki/Likert_scale
Usability
- https://www.ueq-online.org/
- https://en.ryte.com/wiki/System_Usability_Scale/
- https://www.invespcro.com/blog/usability-metrics/
Results
Chapter 7: Results
Statistical Analysis
- Precision and Recall https://en.wikipedia.org/wiki/Precision_and_recall Precision, Recall, Accuracy, F1 Score
- Inter-rater reliability https://en.wikipedia.org/wiki/Inter-rater_reliability
- Fleiss' kappa https://en.wikipedia.org/wiki/Fleiss%27_kappa https://datatab.de/tutorial/fleiss-kappa
- Krippendorff's alpha https://en.wikipedia.org/wiki/Krippendorff%27s_alpha https://real-statistics.com/reliability/interrater-reliability/krippendorffs-alpha/krippendorffs-alpha-basic-concepts/
Threads to validity
Publication
Submission
Presentation
- As a rule of thumb, a slide takes 30 seconds to 2 minutes. If you have a 30-minute presentation, aim for 15 to 30 slides (not counting animation slides)
- Do not exceed the max time, but stay ~10% below it. 30 Minutes mean 25-30 is the ideal time to land.
- Do not include a "thank you for your attention" slide, but close on a conclusion slide, showing everything important to remember on one slide
Practice
Take your presentation for a walk
If you want, you can download your presentation to your phone and just go for a walk, practicing it out loud. Taking this walk outside has the benefit of some fresh air and seeing something else, but speaking aloud might be a problem. If you practice in your room, you flip those up- and downsides, speaking freely without the fresh air. Either should work to give you a feeling of
- how long your segments will take,
- which parts need more dedicated explanation or can be cut entirely without loss, and finally
- confidence when you become less and less reliant on the actual slides to speak freely through your entire talk, since you already did it 10 times.
Publication
- Repository
What then
Lessons Learned: Publication
Dear reader, this is your time to shine! Contribute your own lessons learned, add helpful hints and mistakes made so that others may learn from it. Just click "edit" and add your own subsection!
- ↑ "[D] Overleaf is down?" from 2025-05-13 AOE, https://www.reddit.com/r/MachineLearning/comments/1km8d7p/d_overleaf_is_down/
- ↑ Guarín-Zapata, Nicolás (2022-09-23). "Using MikTex with LaTeX Workshop on Windows". Nicolás' blog. Retrieved 2024-04-26.
- ↑ Shapira, Philip (2024-03-31). "Delving into "delve"". Philip Shapira. Retrieved 2025-01-03.
- ↑ Wikipedia on "Predatory publishing" https://en.wikipedia.org/wiki/Predatory_publishing