Instructions for Writing
Master's Theses

By professor Jacob Palme,
Department of Computer and Systems Sciences
Stockholm University

Latest revision: 5 September, 2013


These instructions apply to master's theses supervised by me. They are not absolute rules, you can choose to do things in different ways if you have valid reasons for doing so.

Scientific structure

A master's thesis at DSV must have a scientific structure. There must be a basis for all statements. This basis can be references to other people's results or your own investigations.

Future sense

Do not write “we will” or other use of future sense. Write in imperfect even if you describe work you have not yet done. Then you need not revise your text for the final version of the thesis.


The “Methods” chapter should describe the methods you have chosen for your thesis, and why you have chosen these methods. Do not write long general chapters about scientific methodology in general.

Literature, References

A thesis should contain a chapter about what other people have done in the same or similar areas. This should be relevant to your own problem and not too general. Do not start a thesis by saying “The Internet started in 1969...”. That is too general.

Suggested format for references is that in the text of your paper, you write for example [Eriksson 2005A] to refer to a paper. In the references list, you then write for example:

[Eriksson 2005A] “Title of the referenced paper” by John Eriksson, Journal of XXX, July, 2005.
Do not use footnotes for references.

References to web pages and Wikipedia articles

Some web pages and Wikipedia article contain very scientifically sound information, others do not. You will have to evaluate each page to decide if it has acceptable quality, before referencing it. In Wikipedia, there is often a coloured box at the top of those articles which are controversial and may not be reliable.

Since web pages or articles in Wikipedia often change, it is best to provide a time-frozen reference to the article you are referencing. You can do this using the Wayback engine or the WebCite, example of a time-frozen reference to Wikipedia.

Funnel model (Swedish: ”Tratt”)

A masters' thesis often has a funnel structure. This means that in the beginning, you describe a wider, general problem area. Then, your own detailed investigation is narrowed to only part of this area. Finally, in the end, you may sometimes widen the scope again and indicate that your results may apply to a wider area than your own investigations.

Chapter structure

Most theses follow approximately this structure:

  1. Problem area.
  2. Literature survey, what other people have done.
  3. Which specific problem you will address.
  4. Why this problem is important.
  5. How you will attack this problem.
  6. How your investigation of this problem progressed.
  7. What results you got.
  8. Conclusions.
  9. Ideas for future work in this area.
  10. Appendices, containing for example technical details which people need not read to understand your results. If the total length of a thesis is more than about 20-40 pages, it is often suitable to move some parts to appendices, so that the main text is not longer than 20-40 pages.


In scientific papers, good practice is to describe any reliability problem. Do not hide such problems. For example, if you make a survey, describe how you selected participants, even if the method was not ideal.

Selection of test persons

This section only applies to papers, in which experiments are done with test persons.

In an ideal study, test persons should be selected by a random sample from a defined population, like for example a random sample of all Swedes. In practice, it is difficult to make such a sample and get people to accept to participate. Because of this, it is generally accepted that you make a sample in less good ways, like for example fellow students or friends or relatives. But you should describe how you made the selection.

Statistical tests of significance, like t-tests, are often used. Theoretically, a t-test is only meaningful if you are using a proper random or stratified sample from a population. But in reality, t-tests are useful even if the sample is not ideal. The t-test will then give some indication on whether your results are based by random fluctuations or not.


If you plan to use a questionnaire, I want to see the questionnaire before you use it.

It is often good to test a questionnaire on a few people, and see how it works, before you finalize the questionnaire to be used for your full study. This testing of the questionnaire with a few people can be done at the same time as I review the questionnaire.

Software development

If you intend to develop software as part of your thesis, you can usually not spend more than half the time on software writing. At least half the time is usually needed for the other parts of the thesis writing.


Submitting something, which someone else has written, and presenting it as your own, is plagiarism and can get you punished. The simple principle to avoid this is to always clearly specify every text you have quoted from someone else. Never use other people's text without indicating the source. The risk is too large. More information about plagiarism.

Time schedule [name=time]

When you have written the introductory chapters about what other people have done, overview of the problem, description of the method you want to do, I would like to see a draft of the thesis at this stage.

When you are ready with the thesis, I will read it and comment on it and maybe suggest changes. This may be needed in several stages. Sometimes there is a queue when many students send manuscripts to me at the same time. Especially in May and December, there are often queues. This may mean that I cannot reply until in one or two weeks.

When I have approved the thesis, then (only for SU students, not KTH) I have to find a separate person to review the thesis. This will cause another delay.

When both I and the reviewer has approved the thesis, it must be submitted two weeks in advance, so that other students can read it.

Thus, you have to accept that these final stages of thesis writing can take many weeks, and does not include full time work all the time for you.

Examples of good theses

Do not look at any thesis written by another student as an example of a good thesis. Many theses are accepted even though they are not very good. In particular, common problems with theses, which are often not corrected in the final version of the thesis are:

  • Many theses are much too long. Think like this: Would I read this thesis if I was not involved with it myself.
  • Detailed transcripts of data: A thesis should not contain complete detailed transcripts of all the data you have compiled. If you absolutely want to include such transcripts, put them into attachments and not into the main text of the thesis.
  • Too much well-known information, like HMI principles, the difference between qualitative and quantitative research, etc. Only include in the thesis what is relevant to this particular thesis. For example, the methods chapter should include the particular methods you have used, reasons why you chose them, maybe reasons why you did not chose other methods. But it should not be a general tutorial on research methodology.

If you want to see good example theses, use those designated by tutors as good theses, not just any theses.

Suitable Thesis Size

Comment: I would recommend a size of a thesis as about 20-60 pages. If the thesis becomes much longer than 20 or 30 pages, it is advisable that the main text is 20-30 pages (with about 500 words/page) and additional text is moved to appendices. The reason for this is that you can seldom expect the audience of your thesis to spend more effort than reading 20-30 pages. Long theses will not be read! Note also that scientific papers in scientific journals and at scientific conferences are usually restricted in size.

Communication with me

I am seldom in my office. If you want to contact me, use e-mail ( or phone. Best time to reach me by phone (08-664 77 48) is usually around 11 a.m. If you want to meet with me, book a time in advance by phone or e-mail.

Do not expect me to contact you. If you do nothing, I will not do anything to remind you.

When writing e-mail to me, begin the subject line of the e-mail message with one of the thee words VIKTIGT or BRÅDSKANDE or [HANDLEDNING]. This tells me that your mail is important, and reduces the risk that the mail is lost as spam. Note the square parenthesises around HANDLEDNING.

If I do not respond to your e-mail within a week, write again. I get so many messages, that I sometimes miss some of them.

If you are two people writing a thesis together

When you send e-mail to me, always include the other thesis author as a CC recipient of your e-mail. In that way, my replies will automatically go to both of you.

Stages of a thesis and reporting to Daisy/LADOK for Stockholm University students

A 30-hp thesis gets points according to the following stages for Stockholm University students:

Stage ECTS Points Results
1 7.5 hp Thesis proposal. A written description of a few pages describing what you aim to do.
2 6 hp Half the thesis ready.
3 9 hp Whole text ready first time.
4 6 hp Revision and final seminar.
5 1.5 hp Opposition and participation on other theses.

Stages of a thesis and reporting to Daisy/LADOK for KTH students

A 30-hp thesis gets points according to the following stages for KTH students:

Stage ECTS Points Results
1 7.5 hp Thesis proposal. A written description of a few pages describing what you aim to do.
2 15 hp Writing the thesis.
3 6 hp Opposition and prescense for other's theses.
4 1.5 hp Final seminar, revision after final seminar.

Getting credit for what you have done

When you feel that you are ready with one of these stages, and want me to enter this into Daisy, tell me explicitly. Otherwise I may forget to do it.

Word processor

If you are using Microsoft Word for writing the thesis, send your drafts to me in Word format (not PDF). You may of course use other editors, and in that case, send in PDF or RTF format to me.


I can supervise theses in Swedish or English. Writing a thesis in English gives you the opportunity to exercise your English skills.

Links to more Information