A mind map is a good study tool for exactly one job: working out how the pieces of a topic relate to each other. It is a bad study tool for memorisation, and most of the disappointment with mind mapping comes from using it for the second thing.
The evidence is reasonably clear on this. Organising material into a structure you built yourself helps. That is the generation effect, and it is one of the better-supported findings in learning research. But looking at the resulting map, however beautiful, is re-reading, and re-reading is close to the weakest study method there is. The map is the thing you make; recall is the thing you practise.
So the method below produces maps that are deliberately incomplete, because completeness is what makes a map unusable for testing yourself.
Build the map from questions, not headings
The default failure is a map that mirrors the textbook: chapter in the middle, section headings as branches, bullet points as leaves. That's a table of contents drawn as a starburst. It took an hour and it taught you the chapter's layout.
Use questions as branch labels instead:
- "What causes X?" rather than "Causes"
- "How does X differ from Y?" rather than "Comparison"
- "When does this stop being true?" rather than "Limitations"
A question branch is testable: you can cover the children and try to answer it. A noun branch is not; you can only look at it and feel familiar, which is the feeling that makes people walk into exams underprepared.
The four-branch skeleton
For nearly any topic, these four get you started, and the fourth is the one that matters:
- Definitions — the terms you must be able to state precisely. Keep this branch short; if it has thirty items, they are not all definitions.
- Mechanisms — how the thing actually works, step by step. This is the branch that should be deepest.
- Relationships — what connects to what, what causes what, what contradicts what. Cross-links, not children.
- Exceptions and edge cases — where the general rule fails. Exam questions live here disproportionately, and this branch is almost always missing from maps made from lecture slides.
The method
1. Map from memory first, then check
Close the book. Build the map from what you can recall, in twenty minutes, with no source material. It will be full of holes.
Then open the source and fill the gaps in a different colour. That colour is now a map of what you did not know, which is the single most useful piece of information a study session can produce. Most people build the map with the book open and lose that signal entirely.
2. Keep leaves to a few words
A branch label that is a full sentence has already given you the answer, which means you can never test yourself on it again. Three or four words is the right size: enough to identify the idea, not enough to state it.
This is also why exported-from-notes maps rarely work. Notes are sentences; a map made of sentences is notes in a worse shape.
3. Cross-link aggressively
The links between branches carry more exam value than the branches. "This mechanism is the reason for that exception" is precisely the kind of connection a question is built to test, and it is the kind of thing a linear set of notes cannot represent at all.
If your map has no cross-links after an hour, you have built a hierarchy, not a map, and the topic almost certainly has connections you have not noticed yet.
4. Collapse and recall
This is the study session, and it is where the map earns its cost. Collapse every branch to its first level. Try to reproduce what's underneath, out loud or on paper. Expand to check.
Do this three days later, then a week after that. Spaced retrieval against a structure you built yourself is a strong combination, considerably stronger than either the map or the flashcards alone.
Collapsing a branch and expanding it again is the whole study loop, and it is one tap here. Checkboxes with progress that rolls up the branch mean a topic reports how much of it you have actually recalled, and the outline view lets you type a list and get a map when you would rather not drag anything.
What it does5. Convert the weak parts to flashcards
Where the recall step failed, you need repetition rather than structure. Pull those nodes into whatever spaced-repetition tool you use.
Most mind map apps export Markdown or OPML, and both are plain text you can reshape into a deck in a few minutes. This is the division of labour that works: map for understanding, cards for memorising. Trying to make one tool do both produces a map too dense to navigate and cards with no context.
Maps for different kinds of material
Processes and sequences need a linear or timeline layout, not a radial one. A process drawn as a starburst loses the one property that mattered: the order.
Comparisons want a matrix layout, or one branch per criterion with the options as children. The instinct is one branch per option, and it is wrong: it makes comparing them require holding both branches in your head at once.
Cases and judgments take fixed branches: facts, holding, reasoning, dissent. The reasoning branch is the one exam answers come from.
Formulae and derivations: the map holds when to use each formula and what each term means. The derivations themselves belong on paper; a mind map is a poor place for anything you need to write in sequence.
What mind mapping is not good for
Rote lists. Vocabulary. Anything where the answer is a fact with no structure around it: dates, names, drug dosages, irregular verbs. Spaced repetition beats mapping for all of it, and it isn't close.
Nor is a map a substitute for practice questions. The map tells you how a topic fits together; a past paper tells you what the examiner does with it. Where they disagree, the past paper is right and your map needs a new branch.
Group study, briefly
One map, built together, is a good use of an hour, for a reason that has little to do with the map: disagreements about where a node belongs are disagreements about the material, and they surface immediately. Silence about where something goes usually means nobody understands it.
Split by branch, merge at the end, then have each person recall a branch they did not write.