How to Take Notes for a Thesis or Dissertation

The failure mode is predictable and it has a shape. Eighteen months of reading, four hundred sources, a folder of highlighted PDFs, and no ability to write a paragraph. Not because you do not know the field, but because your notes are organised by what you read rather than by what you think.

Everything below is aimed at that one problem.

The two organisations, and why they conflict

Reading notes naturally organise by source: one file per paper, in the order you read them. Writing needs them organised by claim: everything relevant to "the measurement problem in this subfield" in one place, regardless of which twelve papers it came from.

You cannot write from the first arrangement. So the work of a thesis is, in large part, the conversion between them, and it is far cheaper to do that conversion continuously than in one panicked term.

Note type Organised by Written when Feeds
Source note The paper On reading The literature review, and nothing else
Idea note The claim When you have a thought Chapters
Chapter outline Your argument From month three onward The thesis
Method and decision log Chronology Continuously The methods chapter and your viva

Four documents. The idea notes are the ones that get skipped and the ones that produce the thesis.

Source notes: short, structured, same every time

One note per source, written the day you read it, in a fixed shape. Consistency matters more than depth here, because you will be reading across these later and irregular notes cannot be read across.

A shape that works:

  • Citation key (from your reference manager, never retyped). See citation management.
  • The claim the paper makes, in one sentence, in your words.
  • Method and sample, briefly. This is what you will need when you compare studies.
  • What is useful to me, which is the actual note.
  • What is weak, because your contribution usually lives in someone else's limitation.
  • Two or three exact quotes with page numbers. Quotes cannot be reconstructed later and getting them wrong is a serious error.

Resist summarising the whole paper. You are not writing an abstract; the abstract already exists.

Idea notes: the part that becomes a thesis

Every time you have a thought that is yours, write it as a separate note with a title that states the claim. "Longitudinal designs in this literature almost never survive attrition past wave three" is a note. "Attrition" is a folder.

These accumulate slowly and then become the argument. The discipline is from the Zettelkasten method, and the specific distinction between reading notes and your own is covered in literature notes and permanent notes.

Two rules make them usable:

One claim per note. A note containing four ideas can only be filed in one place and will only be found by one of the four routes. See atomic notes.

Link them to each other as you go. The links are what makes a chapter appear later. Six months in you will find that eleven notes are all circling the same problem, and that is a chapter you did not know you had. Backlinks and maps of content both help.

The decision log nobody tells you to keep

Keep a dated, append-only file of every methodological choice and why you made it. Why this sample, why that exclusion criterion, why you dropped the second measure, what the supervisor said in March.

You will need this three times: writing the methods chapter, responding to reviewers, and in the viva when someone asks why you did something you decided two years ago for a reason you have completely forgotten. It costs two minutes a week. See how to keep a work journal.

Starting the outline absurdly early

Write a chapter outline in month three, when you know almost nothing. It will be wrong. That is not a problem, because its job is not to be right, it is to give your reading somewhere to go.

A note that has a home in an outline is a note you will use. A note filed under "interesting" is a note you will not. Revise the outline every month and let it be visibly wrong in the meantime.

This is also the antidote to reading forever. An outline makes gaps visible, and a visible gap converts vague anxiety into a specific search.

Writing from the notes

When it is time to write, you should be assembling rather than composing. Pull the idea notes relevant to a section, arrange them in an order, and write the connective tissue.

If that feels impossible, the diagnosis is almost always that the notes are summaries of other people rather than statements of your own. Summaries cannot be assembled into an argument because they have no position. See writing from your notes.

Practical constraints for a multi-year project

  • Choose a format you can still read in five years. Theses outlive apps. File over app.
  • Back up in three places, one of them off-site. This is the one piece of advice everybody agrees on and half of postgraduates ignore until the week it matters. How to back up your notes.
  • Do not migrate systems mid-project. The perfect app will appear in year two. Ignore it.
  • Keep data separate from notes, with the analysis reproducible from the raw file.

The weekly habit that carries it

One hour a week: convert this week's source notes into idea notes, link the new ones to old ones, update the outline, and write down what you are confused about. That single hour is the difference between having read a great deal and having a thesis.

More in how to take research notes, how to read a research paper, and the weekly review.

More in How-to Guides

62 more notes on this branch.

Clair Mind connects your own notes exactly like this — automatically, privately, on your iPhone. Get the app →